The EFSA Journal (2005) 246, 1-110

Transcription

1 The EFSA Journal (00) 46, 11 pinion of the Scientific Panel on Food Additives, Flavourings, Processing Aids and Materials in contact with Food (AFC) on a request from the Commission related to Flavouring Group Evaluation : Aliphatic primary and secondary saturated and unsaturated alcohols, aldehydes, acetals, carboxylic acids and esters containing an additional oxygenated functional group and lactones from chemical groups 9, 13 and 30 (Commission Regulation (EC) No 16/000 of 18 July 000) QUESTIN N EFSAQ00313 Adopted 8 ctober 00 SUMMARY The Scientific 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 Scientific Panel is asked to evaluate 1 flavouring substances in the Flavouring Group Evaluation (FGE.), using the procedure as referred to in the Commission Regulation EC No 16/000. These 1 flavouring substances belong to chemical groups 9, 13 and 30, Annex I of the Commission Regulation EC No 16/000. The present Flavouring Group Evaluation deals with 1 aliphatic primary and secondary saturated and unsaturated alcohols, aldehydes, carboxylic acids, acetals and esters containing an additional oxygenated functional group and lactones. Twentyseven of the 1 flavouring substances possess one or more chiral centres. In 18 of these cases, 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, but the identity of these substances are specified by their names. Three of the substances possess both a double bond and one or more chiral centres, and stereoisomeric information on these substances is not complete as they have been presented without specific indication that the commercial flavouring substance has dominance of one or the other stereoisomer. Fortyeight of the flavouring substances are classified into structural class I and three are classified into structural class II. Fortythree 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 Scientific Panel as a default used the Maximised Surveyderived Daily Intakes (MSDIs) approach to estimate the per capita intakes of the flavouring substances in Europe. However, when the Scientific 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

2 Flavouring Evaluation Group Aliphatic primary and secondary saturated and unsaturated alcohols, aldehydes, acetals, carboxylic acids and esters containing an additional oxygenated functional group and lactones from chemical groups 9, 13 and 30 The EFSA Journal (00) 46 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 Scientific 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 Scientific Panel to make a more realistic estimate of the intakes of the flavouring substances, the Scientific 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 Scientific Panel decided not to carry out a formal safety assessment using the Procedure. In these cases the Scientific Panel requires more precise data on use and use levels. The Procedure was applied to 0 of the 1 flavouring substances, but was deferred for one substance, 1hydroxypropanone [FLno: ], pending further data on genotoxicity. According to the default MSDI approach, the 0 flavouring substances to which the Procedure has been applied have intakes in Europe from to 1 microgram/capita/day which are below the threshold of concern value for both structural class I (1800 microgram/person/day) and structural class II (40 microgram/person/day) substances. It can be anticipated that, at the estimated levels of intake, the flavouring substances included in the present FGE are generally completely metabolised to innocuous products, many of which are endogenous in humans. For three of the candidate substances it cannot be concluded that they are metabolised to innocuous products. These are: butoxyethanol [FLno: 0.4], where the major metabolite, butoxyacetic acid, has been recognised as responsible for haematotoxic effects induced by butoxyethanol; 1,1,3triethoxypropane [FLno: ], which may be metabolised to 3 ethoxypropanoic acid, a substance which has structural similarities to butoxyethanol; and finally, ethyl acetylbutyrate [FLno: 09.84], whose hydrolysis gives rise to acetylbutyric acid, which shows some structural similarities to valproic acid, a known teratogenic compound. For one of the substances [FLno: 09.84] an adequate margin of safety could be established compared to the intake from use as flavouring substances based on the MSDI approach. For two of the substances [FLno: 0.4 and ] no NAEL was available and no margin of safety could be established. n the basis of the data available it is concluded that there is no indication that the candidate substances in the present Flavouring Group Evaluation possess genotoxic potential with the exception of 1hydroxypropanone [FLno: ] which shows positive results in in vitro gene mutation studies in bacteria. It was noted that where toxicity data were available they were consistent with the conclusions in the present Flavouring Group Evaluation using the Procedure. It was considered that on the basis of the default MSDI approach 48 of the 0 flavouring substances, to which the Procedure could be applied, would not give rise to safety concerns at the estimated levels of intake arising from their use as flavouring substances. For butoxyethanol [FLno: 0.4] and 1,1,3triethoxypropane [FLno: ] additional toxicity data are required. When the estimated intakes were based on the mtamdi they ranged from 1600 to 4700 microgram/person/day for the 47 flavouring substances from structural class I for which the Procedure could be applied. Thus the intakes were above the threshold of concern for structural class I of 1800 microgram/person/day except for one of the flavouring substances [FLno: 0.149]. This substance is also expected to be metabolised to innocuous products. The estimated intake of

3 Flavouring Evaluation Group Aliphatic primary and secondary saturated and unsaturated alcohols, aldehydes, acetals, carboxylic acids and esters containing an additional oxygenated functional group and lactones from chemical groups 9, 13 and 30 The EFSA Journal (00) 46 the three flavouring substances assigned to structural class II, based on the mtamdi is 3700 microgram/person/day for each, which is above the threshold of concern for structural class II of 40 microgram/person/day. Thus, for 49 flavouring substances in this opinion the intakes, estimated on the basis of the mtamdi, exceed the relevant threshold for their structural class. Therefore, for these 49 substances more reliable exposure data are required. n the basis of such additional data, these flavouring substances should be reconsidered along the steps of the Procedure. Furthermore, additional data are required for two of the substances [FLno: 0.4 and ] and might become necessary for the remaining 47 substances. 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 0 flavouring substances which could be evaluated through the Procedure, except that information on stereoisomerism is missing for 1 of the substances. Thus, the final evaluation of the materials of commerce cannot be performed for these 1 substances, pending further information. KEYWRDS Flavourings, safety, lactones, saturated, unsaturated, primary, secondary, alcohols, aldehydes, acids, acetals, esters, additional oxygenated functional group. 3

4 Flavouring Evaluation Group Aliphatic primary and secondary saturated and unsaturated alcohols, aldehydes, acetals, carboxylic acids and esters containing an additional oxygenated functional group and lactones from chemical groups 9, 13 and 30 The EFSA Journal (00) 46 TABLE F CNTENTS Summary... 1 Keywords... 3 Background... Terms of Reference... Assessment Presentation of the Substances in Flavouring Group Evaluation (FGE.) Description Stereoisomers Natural ccurrence in Food Specifications Intake data Estimated Daily per Capita Intake (MSDI Approach) Intake Estimated on the Basis of the Modified TAMDI (mtamdi) Absorption, Distribution, Metabolism, and Elimination.... Application of the Procedure for the Safety Evaluation of Flavouring Substances Comparison of the Intake Estimations based on the MSDI Approach and the mtamdi Approach Considerations of Combined Intakes from Use as Flavouring Substances Toxicity Acute Toxicity Subacute, Subchronic, Chronic and Carcinogenicity Studies Developmental / Reproductive Toxicity Studies Genotoxicity Studies Conclusions Table 1: Specification Summary of the Substances in the Flavouring Group Evaluation... Table a: Summary of Safety Evaluation Applying the Procedure (based on intakes calculated by the MSDI approach)... 9 Table b: Evaluation Status of Hydrolysis Products of Candidate Esters Table 3: Supporting Substances Summary Annex I: Procedure for the Safety Evaluation... 3 Annex II: Use Levels / mtamdi... Annex III: Metabolism Annex IV: Toxicity References:... 9 Scientific Panel Members... 1 Acknowledgement

5 Flavouring Evaluation Group Aliphatic primary and secondary saturated and unsaturated alcohols, aldehydes, acetals, carboxylic acids and esters containing an additional oxygenated functional group and lactones from chemical groups 9, 13 and 30 The EFSA Journal (00) 46 BACKGRUND Regulation (EC) No 3/96 of the European Parliament and the Council (EC, 1996) lays down a procedure for the establishment of a list of flavouring substances, the use of which will be authorised to the exclusion of all others in the EU. In application of that Regulation, a register of flavouring substances used in or on foodstuffs in the Member States was adopted by Commission Decision 1999/17/EC (EC, 1999a), as last amended by Commission Decision 00/389/EC (EC, 00). Each flavouring substance is attributed a FLAVISnumber (FLnumber) and all substances are divided into 34 chemical groups. Substances within a group should have some metabolic and biological behaviour in common. Substances which are listed in the register are to be evaluated according to the evaluation programme laid down in Commission Regulation (EC) No 16/000 (EC, 000) which is broadly based on the opinion of the Scientific Committee on Food (SCF, 1999). For the submission of data by the manufacturer, deadlines have been established by Commission Regulation (EC) No 6/00 (EC, 00b). After the completion of the evaluation programme the positive list of flavouring substances for use in or on foods in the EU shall be adopted (Article (1) of Regulation (EC) No 3/96) (EC, 1996). TERMS F REFERENCE EFSA is requested to carry out a risk assessment on flavouring substances prior to their authorisation and inclusion in a positive list according to Commission Regulation (EC) No 16/000 (EC, 000). ASSESSMENT 1. Presentation of the Substances in Flavouring Group Evaluation (FGE.) 1.1. Description The present Flavouring Group Evaluation, using the procedure as referred to in the Commission Regulation EC No 16/000 (EC, 000) (The Procedure shown in schematic form in Annex I), deals with 1 alcohols, aldehydes, acetals, carboxylic acids and esters containing an additional oxygenated functional group and lactones from chemical groups 9, 13 and 30, Annex I of Commission Regulation (EC) No 16/000 (EC, 000). The 1 flavouring substances (candidate substances) under consideration are listed in Table 1, as well as their chemical name, FLAVIS ( FL), Chemical Abstract Service (CAS), Council of Europe (CoE) and Flavor and Extract Manufactures Association (FEMA) numbers, structure and specifications. Eleven candidate substances are aliphatic lactones [FLno:.038,.039,.043,.04,.047,.048,.049,.0,.0,.08 and.068]; twentynine candidate substances are esters or diesters [FLno: , , , 09.0, 09.8, 09.80, 09.90, , 09.66, 09.69, , , , , 09.81, 09.84, 09.83, , and ]; two candidate substances are acetals [FLno: and ]; one candidate substance is an alphahydroxyacid [FLno: ]; one candidate substance is a betaketoacid [FLno: 08.03]; one

6 Flavouring Evaluation Group Aliphatic primary and secondary saturated and unsaturated alcohols, aldehydes, acetals, carboxylic acids and esters containing an additional oxygenated functional group and lactones from chemical groups 9, 13 and 30 The EFSA Journal (00) 46 candidate substance is an alkoxyalcohol [FLno: 0.4]; two candidate substances are diols [FLno: 0.13 and 0.198]; one candidate substance is a dialdehyde [FLno: 0.149] and three candidate substances are aliphatic dicarboxylic acids [FLno: 08.08, and 08.3]. The 1 candidate substances are structurally related to 34 aliphatic lactones (supporting substances) evaluated at the 49 th JECFA meeting (JECFA, 1998a) and to 47 aliphatic primary alcohols, aldehydes, carboxylic acids, acetals, and esters containing additional oxygenated functional groups evaluated at the 3 rd JECFA meeting (JECFA, 000c). These supporting substances are listed in Table 3, together with their evaluation status. The hydrolysis products of candidate esters and acetals as well as their evaluation status are listed in Table b. 1.. Stereoisomers It is recognised that geometrical and optical isomers of substances may have different properties. Their flavour may be different, they may have different chemical properties resulting in possible variation of their absorption, distribution, metabolism, elimination and toxicity. Thus information must be provided on the configuration of the flavouring substance, i.e. whether it is one of the geometrical/optical isomers, or a defined mixture of stereoisomers. The available specifications of purity will be considered in order to determine whether the safety evaluation carried out for candidate substances for which stereoisomers may exist can be applied to the material of commerce. Flavouring substances with different configurations should have individual chemical names and codes (CAS number, FLAVIS number etc.). Twentyseven of the substances possess one or more chiral centres [FLno: 0.13, 0.198, , , , , , , 09.0, 09.80, 09.90, , 09.69, , , , 09.81, 09.84, 09.83, 09.86, ,.04,.048,.049,.0,.08 and.068]. In 18 of these cases, the substances have been presented without any indication whether the commercial flavouring substance was dominated by one enantiomer or the other. Due to the presence and the position of a double bond, five of the 1 substances can exist as geometrical isomers. The stereoisomers of two of these substances [FLno: and 09.31] are specified by their names. The remaining three of the substances which can exist as geometrical isomers also possess one or more asymmetric centres [FLno:.038,.039 and.043]. Stereoisomeric information on these substances is not complete as they have been presented without specific indication that the commercial flavouring substance has dominance of one or the other stereoisomer Natural ccurrence in Food Fortythree of the 1 flavouring substances have been reported to occur in one or more of the following food items: fruits (apple, pineapple, melon, guava, banana, starfruit, papaya, raspberry, mango, plum), juice, butter, meat, cheese, skimmed milk powder, green tea, coffee, beer, wine, and whisky. Quantitative data on the natural occurrence in food have been reported for thirtyfour of the candidate substances. These reports include among others: ctane1,3diol [FLno: 0.198]: up to 1 mg/kg in apple and up to 9.1 mg/kg in apple juice. Hexadecano1,lactone [FLno:.049]: up to.6 mg/kg in butter and up to 1.3 mg/kg in heated lamb and mutton fat. 6

7 Flavouring Evaluation Group Aliphatic primary and secondary saturated and unsaturated alcohols, aldehydes, acetals, carboxylic acids and esters containing an additional oxygenated functional group and lactones from chemical groups 9, 13 and 30 The EFSA Journal (00) 46 Hexadecano1,4lactone [FLno:.048]: up to 16.7 mg/kg in heated butter. Butoxyethan1ol [FLno: 0.4]: 0.0 mg/kg in mozzarella cheese. Hexadecano1,16lactone [FLno:.047]: mg/kg in skimmed milk powder. Heptano1,lactone [FLno:.04]: up to 0.4 mg/kg in green tea. 1Hydroxypropanone [FLno: ]: up to 4 mg/kg in coffee. 1,1,3Triethoxypropane [FLno: ]: up to 3 mg/kg in pear brandy and less than 0.8 mg/kg in wisky. Nonanedioic acid [FLno: 08.3]: up to1. mg/kg in beer. Propyl lactate [FLno: 09.81]: trace amount in white wine. Isobutyl lactate [FLno: 09.90]: 0 mg/kg in port wine. Eight of the substances have not been reported in any food items according to TN (TN, 000). These eight substances are: methyl lactate [FLno: ], ethyl butyryl lactate [FLno: 09.0], pentadecano1,14lactone [FLno:.068], ethyl acetylbutyrate [FLno: 09.84], isopropyl 4 oxopentanoate [FLno: ], ethyl 3acetohexanoate [FLno: 09.83], methyl hydroxydecanoate [FLno: ] and di(methylbutyl) malate [FLno: ].. Specifications Purity criteria for the 1 substances have been provided by the flavouring industry (EFFA, 003c) (Table 1). Judged against the requirements in Annex II of Commission Regulation 16/000 (EC, 000), this information is adequate for all the 1 substances. However, information on geometrical stereoisomerism/chirality is needed for 1 of the candidate substances (see section 1. and Table 1). 3. Intake data Annual production volumes of the flavouring substances as surveyed by the Industry can be used to calculate the Maximized SurveyDerived Daily Intake (MSDI) by assuming that the production figure only represents 60 % of the use in food due to underreporting and that % of the total EU population are consumers (SCF, 1999). However, the Panel noted that due to yeartoyear variability in production volumes, to uncertainties in the underreporting correction factor and to uncertainties in the percentage of consumers, the reliability of intake estimates on the basis of the MSDIapproach is difficult to assess. The Panel also noted that in contrast to the generally low per capita intake figures estimated on the basis of this MSDIapproach, in some cases the regular consumption of products flavoured at use levels reported by the Flavour Industry in the submissions would result in much higher intakes. In such cases, the human exposure thresholds below which exposures are not considered to present a safety concern might be exceeded. 7

8 Flavouring Evaluation Group Aliphatic primary and secondary saturated and unsaturated alcohols, aldehydes, acetals, carboxylic acids and esters containing an additional oxygenated functional group and lactones from chemical groups 9, 13 and 30 The EFSA Journal (00) 46 Considering that the MSDI model may underestimate the intake of flavouring substances by certain groups of consumers, the SCF recommended also taking into account the results of other intake assessments (SCF, 1999). ne of the alternatives is the Theoretical Added Maximum Daily Intake (TAMDI)approach which is calculated on the basis of standard portions and upper use levels (SCF, 199) for flavourable beverages and foods in general, with exceptional levels for particular foods. This method is regarded as a conservative estimate of the actual intake in most consumers because it is based on the assumption that the consumer regularly eats and drinks several food products containing the same flavouring substance at the upper use level. ne option to modify the TAMDIapproach is to base the calculation on normal rather than upper use levels of the flavouring substances. This modified approach is less conservative (e.g., it may underestimate the intake of consumers being loyal to products flavoured at the maximum use levels reported (EC, 000). However, it is considered as a suitable tool to screen and prioritise the flavouring substances according to the need for refined intake data (EFSA, 004a) Estimated Daily per Capita Intake (MSDI Approach) The Maximised SurveyDerived Daily Intake (MSDI (SCF, 1999)) data are derived from surveys on annual production volumes in Europe. These surveys were conducted in 199 by the International rganization of the Flavour Industry, in which flavour manufacturers reported the total amount of each flavouring substance incorporated into food sold in the EU during the previous year (IFI, 199). The intake approach does not consider the possible natural occurrence in food. Average per capita intake (MSDI) is estimated on the assumption that the amount added to food is consumed by % of the population 1 (Eurostat, 1998). This is derived for candidate substances from estimates of annual volume of production provided by Industry and incorporates a correction factor of 0.6 to allow for incomplete reporting (60%) in the Industry surveys (SCF, 1999). The total annual volumes of production of the 1 candidate substances from use as flavouring substances in Europe has been reported to be approximately 300 kg (EFFA, 003d; EFFA, 000c). For the 81 supporting substances the annual volume of production is kg (JECFA, 1999b; JECFA, 000b). n the basis of the annual volumes of production reported for the 1 candidate substances, the daily per capita intakes for each of these flavourings have been estimated (Table a). Nearly 90 % of the total annual volumes of production for the candidate substances is accounted for by five of these flavourings: secbutyl lactate [FLno: ], diethyl maleate [FLno: 09.31], ethyl acetoxypropionate [FLno: ], isobutyl lactate [FLno: 09.90] and isopentyl lactate [FLno: ]. The estimated daily per capita intake of secbutyl lactate from use as a flavouring substance is 3.7 microgram, that of diethyl maleate is 1 microgram, that of ethyl acetoxypropionate is 4.9 microgram, that of isobutyl lactate is 3.7 microgram, and that of isopentyl lactate is 7. microgram. The daily per capita intakes for each of the remaining substances are less than 1. microgram (Table a). 1 EU figure 37 millions. This figure relates to EU population at the time for which production data are available, and is consistent (comparable) with evaluations conducted prior to the enlargement of the EU. No production data are available for the enlarged EU. 8

9 Flavouring Evaluation Group The EFSA Journal (00) 46 Aliphatic primary and secondary saturated and unsaturated alcohols, aldehydes, acetals, carboxylic acids and esters containing an additional oxygenated functional group and lactones from chemical groups 9, 13 and Intake Estimated on the Basis of the Modified TAMDI (mtamdi) The method for calculation of modified Theoretical Added Maximum Daily Intake (mtamdi) values is based on the approach used by SCF up to 199 (SCF, 199). The assumption is that a person may consume a certain amount of flavourable foods and beverages per day. For the 1 candidate substances information on food categories and normal and maximum use levels,3 were submitted by the Flavour Industry (EFFA, 003c; EFFA, 001a). The 1 candidate substances are used in flavoured food products divided into the food categories, outlined in Annex III of the Commission Regulation 16/000 (EC, 000), as shown in Table 3.1. For the present calculation of mtamdi, the reported normal use levels were used. In the case where different use levels were reported for different food categories the highest reported normal use level was used. Table 3.1 Use of Candidate Substances Food category Description Flavourings used Category 1 Dairy products, excluding products of category All 1 Category Fats and oils, and fat emulsions (type waterinoil) All 1 Category 3 Edible ices, including sherbet and sorbet All 1 Category 4.1 Processed fruits All 1 Category 4. Processed vegetables (incl. mushrooms & fungi, roots & tubers, pulses and legumes), and nuts & seeds None Category Confectionery All 1 Category 6 Cereals and cereal products, incl. flours & starches from roots & tubers, pulses & legumes, excluding bakery All 1 Category 7 Bakery wares All 1 Category 8 Meat and meat products, including poultry and game All 1 Category 9 Fish and fish products, including molluscs, crustaceans and echinoderms All except [FLno: ] Category Eggs and egg products None Category 11 Sweeteners, including honey None Category 1 Salts, spices, soups, sauces, salads, protein products etc. All except [FLno: ] Category 13 Foodstuffs intended for particular nutritional uses. All except [FLno: ] Category 14.1 Nonalcoholic ("soft") beverages, excl. dairy products All 1 Category 14. Alcoholic beverages, incl. alcoholfree and lowalcoholic counterparts None Category 1 Readytoeat savouries All 1 Category 16 Composite foods (e.g. casseroles, meat pies, mincemeat) foods that could not be placed in categories 1 1 All 1 According to the flavour industry the normal use levels for the 1 candidate substances are in the range of 1 0 mg/kg food, and the maximum use levels are in the range of 0 mg/kg (EFFA, 003c; EFFA, 001a). Normal use is defined as the average of reported usages and maximum use is defined as the 9th percentile of reported usages (EFFA, 00i) 3 The normal and maximum use levels in different food categories (EC, 000) have been extrapolated from figures derived from 1 model flavouring substances (EFFA, 004e). 9

10 Flavouring Evaluation Group Aliphatic primary and secondary saturated and unsaturated alcohols, aldehydes, acetals, carboxylic acids and esters containing an additional oxygenated functional group and lactones from chemical groups 9, 13 and 30 The EFSA Journal (00) 46 The mtamdi values for the 1 candidate substances, for which use levels have been reported, from structural class I range from 1600 to 4700 microgram/person/day, for the three candidate substances from structural class II the mtamdi is 3700 microgram/person/day for each. For detailed information on use levels and intake estimations based on the mtamdi approach, see Section 6 and Annex II. 4. Absorption, Distribution, Metabolism, and Elimination A more detailed description of the metabolism of the candidate substances in this FGE is given in Annex III. In general, lactones are formed by acidcatalysed intramolecular cyclisation of gamma or deltahydroxycarboxylic acids. In an aqueous environment, a phdependent equilibrium is established between the openchain hydroxycarboxylate anion and the lactone ring. In basic and neutral media, such as blood, the openchain hydroxycarboxylate anion is favoured while in acidic media, such as gastric juice and urine, the lactone ring is favoured. Enzymes, such as lactonase, may catalyze the hydrolysis reaction, but for simple saturated lactones, the ringopening reaction and reverse cyclication are in equilibrium, mainly controlled by ph conditions. Both the aliphatic lactones and the ringopened hydroxycarboxylic acids can be absorbed from the gastrointestinal tract. However, the simple lactones with low molecular weight being uncharged may cross the cell membrane more easily than the acidic form, which penetrates the cells as a weak electrolyte. The hydroxycarboxylic acid obtained from lactone hydrolysis enters the fatty acid pathway and undergoes alpha or betaoxidation and cleavage to form acetyl CoA and a chainshortened carboxylic acid. The carboxylic acid is then reduced by carbon fragments until either acetyl CoA or propionyl CoA is produced. These fragments are then metabolised in the citric acid cycle. Mono and diesters included in the present FGE are expected to undergo hydrolysis in humans to yield their corresponding alcohol (saturated linear or branchedchain aliphatic primary alcohols, or branchedchain hydroxy or keto alcohols) and acid components (i.e. alpha, beta or gammaketo or hydroxy acids; or simple aliphatic acids, diacids or triacids), which would be further metabolised and excreted. It has to be noted that the acetyl butyric acid, formed as one of the hydrolysis products of the candidate substance ethyl acetylbutyrate [FLno: 09.84], has some structural similarities to valproic acid, which, together with a number of its derivatives, has been recognised as teratogenic in rodents and in humans (Nau & Löscher, 1986; Samren et al., 1997; Kaneko et al., 1999). Although it can be predicted that acetylbutyric acid is further metabolised through the usual pathways of detoxication for carboxylic acids (i.e. mainly via glucuronidation reaction), the structural similarity with valproic acid does not allow the prediction that ethyl acetylbutyrate [FLno: 09.84] is metabolised only to innocous products. The presence of a second oxygenated functional group has little if any effect on hydrolysis of these esters. The most probable metabolic reactions of the hydrolysis products are: oxidation of alcohols to aldehydes and acids; conjugation of alcohols and acids to glucuronides and sulphates; beta and omegaoxidation of carboxylic acids. Betaketo acids and derivatives like acetoacetic acid undergo ready decarboxylation. Along with alphaketo and alphahydroxyacids, they yield breakdown products, which are incorporated into normal biochemical pathways. The gammaketo acids and related substances may undergo complete or partial betaoxidation to yield metabolites that are eliminated in the urine. megasubstituted derivatives are readily oxidised and/or excreted in the urine. Simple aliphatic di and tricarboxylic acids participate in the tricarboxylic acid cycle.

11 Flavouring Evaluation Group Aliphatic primary and secondary saturated and unsaturated alcohols, aldehydes, acetals, carboxylic acids and esters containing an additional oxygenated functional group and lactones from chemical groups 9, 13 and 30 The EFSA Journal (00) 46 Two candidate substances [FLno: and ] are acetals, which may be expected to undergo acid catalysed hydrolysis in the gastric environment to yield their component aldehydes and alcohols prior to absorption. nce hydrolysed, the component alcohols and aldehydes are expected to be metabolised primarily through the above mentioned common routes of biotransformations and excreted. The linear and branchedchain aliphatic primary alcohol components of candidate substances that are simple aliphatic di and tricarboxylic acid esters would be oxidised in the presence of alcohol dehydrogenase to their corresponding aldehydes which, in turn, would be oxidised to their corresponding carboxylic acids. The two diols [FLno: 0.13 and 0.198] may be anticipated to participate in the same routes of biotransformation. It may be anticipated that glutaraldehyde [FLno: 0.149] is biotransformed through the common pathways of detoxication of aldehydes to innocuous products. Among the candidate substances, butoxyethanol [FLno: 0.4] an alkoxyalcohol is mainly metabolised to butoxyacetic acid, which has been identified as the metabolite responsible for the haemolysis of red blood cells induced by butoxyethanol. In summary, it can be anticipated that primary and secondary aliphatic saturated or unsaturated alcohols, aldehydes, carboxylic acids, acetals and esters with a second oxygenated functional group and aliphatic lactones included in the present FGE are generally metabolised to innocuous products (many of which are endogenous in humans), at the estimated level of intake as flavouring substances. The consideration on the actual levels of intake becomes particularly relevant for one candidate substance, diethyl maleate [FLno: 09.31]; since when administered at high doses, it is able to induce severe GSH depletion, due to its prompt metabolism to GSHconjugates. This may also be the case for the structually related diethyl fumarate [FLno: 09.30]. For three of the candidate substances it cannot be concluded that they are metabolised to innocuous products. These are: butoxyethan1ol [FLno: 0.4], where the major metabolite, butoxyacetic acid, has been recognised as responsible for haematotoxic effects induced by butoxyethanol; 1,1,3triethoxypropane [FLno: ], which may be metabolised to 3ethoxypropanoic acid, a substance which has structural similarities to butoxyethanol; and finally, ethyl acetylbutyrate [FLno: 09.84], whose hydrolysis gives rise to acetylbutyric acid, which shows some structural similarities to valproic acid, a known teratogenic compound.. Application of the Procedure for the Safety Evaluation of Flavouring Substances The application of the Procedure is based on intakes estimated on the basis of the MSDI approach. Where, the mtamdi approach indicates that the intake of a flavouring substance might exceed its corresponding threshold of concern, a formal safety assessment is not carried out using the Procedure. In these cases the Panel requires more precise data on use and use levels. For comparison of the intake estimations based on the MSDI approach and the mtamdi approach, see Section 6. ne of the candidate substances, 1hydroxypropanone [FLno: ] was found to be genotoxic in vitro in bacteria. Therefore, in the absence of further genotoxicity data, the Panel concluded that the Procedure could not be applied to this substance. 11

12 Flavouring Evaluation Group Aliphatic primary and secondary saturated and unsaturated alcohols, aldehydes, acetals, carboxylic acids and esters containing an additional oxygenated functional group and lactones from chemical groups 9, 13 and 30 The EFSA Journal (00) 46 For the safety evaluation of the remaining 0 candidate substances from chemical groups 9, 13 and 30 the Procedure as outlined in Annex I was applied, based on the MSDI approach. The stepwise evaluations of the 0 substances are summarised in Table. Step 1 Fortyseven of the candidate substances are classified according to the decision tree approach by Cramer et al. (Cramer et al., 1978) into structural class I and three are classified into structural class II [FLno: 0.4, and ]. Step For three of the candidate substances it cannot be concluded that they are metabolised to innocuous products. Butoxyethanol [FLno: 0.4], where the major metabolite, butoxyacetic acid, has been recognised as responsible for haematotoxic effects induced by butoxyethanol, and 1,1,3 triethoxypropane [FLno: ], which may be metabolised to 3ethoxypropanoic acid, a substance which has structural similarities to butoxyethanol; and finally, ethyl acetylbutyrate [FLno: 09.84], whose hydrolysis gives rise to acetylbutyric acid, which shows some structural similarities to valproic acid, a known teratogenic compound. Therefore, these substances are evaluated via the Bside of the Procedure. The evaluation of the remaining 47 candidate substances proceeds via the Aside of the Procedure. Step A3 Step A3 applies to the following 46 candidate substances all from structural class I: [FLno: 0.13, 0.198, 0.149, 08.03, 08.08, , 08.3, , , , 09.0, 09.8, 09.80, 09.90, , 09.66, 09.69, , , , , 09.81, 09.83, , 09.86, ,.038,.039,.043,.04, ,.0,.0,.08 and.068], and one candidate substance from structural class II: [FLno: ]. The 46 candidate substances which have been assigned to structural class I have estimated European daily per capita intakes (MSDI) ranging from to 1 microgram and the one candidate substance which has been assigned to structural class II has an estimated European daily per capita intake of 0.01 microgram (Table 6.1). These intakes are below the thresholds of concern of 1800 microgram/person/day and 40 microgram/person/day for structural class I and structural class II, respectively. Accordingly, these 47 candidate substances do not pose a safety concern when used at estimated levels of intake, based on the MSDI approach, as flavouring substances. Step B3 The MSDI of the candidate substances butoxyethan1ol [FLno: 0.4], 1,1,3triethoxypropane [FLno: ] and ethyl acetylbutyrate [FLno: 09.84], were estimated to be microgram/capita/day for each. Thus, the MSDIvalues of all three candidate substances are below the threshold of concern for their structural classes of 40 microgram/person/day (class II) for [FLno: 0.4 and ] and of 1800 microgram/person/day (class I ) for [FLno: 09.84]. Accordingly, the three substances proceed to step B4 of the Procedure. Step B4 The candidate substance ethyl acetylbutyrate [FLno: 09.84] is expected to be hydrolysed to the corresponding alphaethylated carboxylic acid, acetylbutyric acid and ethanol. No toxicity studies 1

13 Flavouring Evaluation Group Aliphatic primary and secondary saturated and unsaturated alcohols, aldehydes, acetals, carboxylic acids and esters containing an additional oxygenated functional group and lactones from chemical groups 9, 13 and 30 The EFSA Journal (00) 46 that would permit establishment of a NAEL are available for ethyl acetylbutyrate or its hydrolysis product acetylbutyric acid. Acetylbutyric acid is structurally related to ethylhexanol [FLno: 0.08] for which JECFA has established an ADI of 0. mg/kg bw/day (JECFA, 1993b). The estimated daily per capita intake, based on the MSDI approach and expressed in microgram/kg bw/day for the hydrolysis product of the candidate substance ethyl acetylbutyrate (and acetylbutyric acid) is approximately x 6 fold below the ADI value. Further the hydrolysis product, acetylbutyric acid, shows some structural similarities to valproic acid, a known teratogenic compound. If acetylbutyric acid is considered to be as potent as valproic acid (NAEL = 600 mg/day) the margin of safety would be x 8, based on the MSDI of microgram/capita/day. Accordingly, it is concluded that ethyl acetylbutyrate [FLno: 09.84] does not pose a safety concern at the estimated level of intake, based on the MSDI approach. For the candidate substances butoxyethan1ol [FLno: 0.4] and 1,1,3triethoxypropane [FLno: ] the hydrolysis product of which has structural similarities to butoxyethan1ol, a NAEL could not be established in subchronic/chronic toxicity studies with respect to haematoxicity. Therefore the Panel concluded that additional toxicity data are needed for these two substances. 6. Comparison of the Intake Estimations based on the MSDI Approach and the mtamdi Approach The mtamdi for the 47 candidate substances in structural class I, to which the Procedure has been applied, range from 1600 to 4700 microgram/person/day. For 46 of these substances the mtamdi is above the threshold of concern of 1800 microgram/person/day. The mtamdi of the three substances assigned to structural class II, to which the Procedure has been applied, is 3700 microgram/person/day, which is above the threshold of concern for structural class II substances of 40 microgram/person/day. For comparison of the MSDI and mtamdivalues see Table 6.1. For 47 candidate substances further information is required. This would include more reliable intake data and then, if required, additional toxicological data. Table 6.1 Estimated intakes based on the MSDI approach and the mtamdi approach FLno EU Register name MSDI (µg/capita/day) mtamdi (µg/person/day) Structural class Threshold of concern (µg/person/day) 0.13 Butane1,3diol Class I ctane1,3diol Class I Glutaraldehyde Class I Malonic acid Class I Glutaric acid Class I Hydroxy4methylvaleric acid Class I Nonanedioic acid Class I secbutyl lactate Class I Diisopentyl succinate Class I Dibutyl malate Class I Dibutyl succinate Class I Diethyl adipate Class I Diethyl citrate Class I Diethyl fumarate Class I Diethyl maleate Class I

14 Flavouring Evaluation Group Aliphatic primary and secondary saturated and unsaturated alcohols, aldehydes, acetals, carboxylic acids and esters containing an additional oxygenated functional group and lactones from chemical groups 9, 13 and 30 The EFSA Journal (00) 46 Table 6.1 Estimated intakes based on the MSDI approach and the mtamdi approach FLno EU Register name MSDI (µg/capita/day) mtamdi (µg/person/day) Structural class Threshold of concern (µg/person/day) 09.3 Diethyl nonanedioate Class I Diethyl oxalate Class I Diethyl pentanedioate Class I Ethyl acetoxypropionate Class I Ethyl butyryl lactate Class I Dimethyl malonate Class I Hexyl lactate Class I Isobutyl lactate Class I Isopentyl lactate Class I Methyl oxopropionate Class I Methyl 3acetoxyhexanoate Class I Methyl hydroxydecanoate Class I Methyl acetoacetate Class I Methyl lactate Class I Pentyl lactate Class I Propyl lactate Class I Ethyl 3acetohexanoate Class I isopropyl 4oxopentanoate Class I Ethyl 3acetoxy octanoate Class I Di(methylbutyl) malate Class I Dec7eno1,4lactone Class I cisdec7eno1,4lactone Class I ,7Dimethylocta(trans),7dieno1,4lactone Class I Heptano1,lactone Class I Hexadecano1,16lactone Class I Hexadecano1,4lactone Class I Hexadecano1,lactone Class I Methylnonano1,4lactone Class I Pentano1,lactone Class I Tridecano1,lactone Class I Pentadecano1,14lactone Class I Ethyl acetylbutyrate Class I Hydroxypropanone Class I Hydroxymethylmethyl1,3dioxolane Class II Butoxyethan1ol Class II ,1,3Triethoxypropane Class II Considerations of Combined Intakes from Use as Flavouring Substances Because of structural similarities of candidate and supporting substances, it can be anticipated that many of the flavourings are metabolised through the same metabolic pathways and that the metabolites may affect the same target organs. Further, in case of combined exposure to structurally related flavourings, the pathways could be overloaded. Therefore, combined intake should be considered. As flavourings not included in this Flavouring Group Evaluation may also be metabolised through the same pathways, the combined intake estimates presented here are only preliminary. Currently, the combined intake estimates are only based on MSDI exposure estimates, although it is recognised that this may lead to underestimation of exposure. After completion of all FGEs, this issue should be readdressed. The total estimated combined daily per capita intake is estimated by summing the MSDI for individual substances. 14

15 Flavouring Evaluation Group Aliphatic primary and secondary saturated and unsaturated alcohols, aldehydes, acetals, carboxylic acids and esters containing an additional oxygenated functional group and lactones from chemical groups 9, 13 and 30 The EFSA Journal (00) 46 n the basis of the reported annual production volumes in Europe (EFFA, 003d), the combined estimated daily per capita intake as flavourings of the 48 candidate flavouring substances assigned to structural class I is 40 microgram, which does not exceed the threshold of concern for a substance belonging to structural class I of 1800 microgram/person/day. n the basis of the reported annual production volumes in Europe (EFFA, 003d) the estimated daily per capita intake as flavourings of the three candidate flavouring substances assigned to structural class II is 0.01 microgram, which does not exceed the threshold of concern for a substance belonging to structural class II of 40 microgram/person/day. The 11 candidate lactones are structurally related to 9 supporting lactones from structural class I, for which the combined intake based on the MSDI approach is 0000 microgram/capita/day. The supporting substances were evaluated by JECFA at the 49 th meeting, where it was noted that although the combined intake exceeds the threshold for the structural class, the lactones are expected to be hydrolysed and completely metabolised to innocuous products at the estimated level of intake as flavouring substances, and would not give rise to perturbations outside the physiological range. The Panel agreed with this view and concluded that the combined intake of about 4 microgram/capita/day for the candidate lactones is negligible compared to the combined intake of 0000 microgram/capita/day of the supporting lactones. Likewise 37 candidate substances are structurally related to 46 supporting aliphatic primary alcohols and related substances containing an additional oxygenated functional group from structural class I. The combined intake of these supporting substances amounts to 4000 microgram/capita/day based on the MSDI approach. These substances were evaluated at the 3 rd JECFA meeting, where it was also noted that the substances are expected to be efficiently metabolised to innocuous products and would not give rise to perturbations outside the physiological range. The Panel agreed with this view and concluded that the contribution from the combined intake of the candidate substances of 3 microgram/capita/day would not alter the JECFA conclusion based on a combined intake of 4000 microgram/capita/day. 8. Toxicity 8.1. Acute Toxicity Data are available for sixteen of the candidate substances (Annex IV, Table IV.1). For the majority of candidate substances, oral LD 0 values, in mice or rats, varied from 0 mg/kg up to >000 mg/kg bw. For butane1,3diol [FLno: 0.13] and octane1,3diol [FLno: 0.198] LD 0 values between 0 g/kg bw/day and approximately 30 g/kg bw/day are reported (Annex IV, Table IV.1). Forty seven (out of 81) supporting substances were tested for acute toxicity in mice and/or rats (Annex IV, Table IV.1). For the majority of the supporting substances, oral LD 0 values, in mice or rats, varied from 1300 mg/kg up to 1800 mg/kg bw. For diethyl sebacate [FLno: 09.47] and tributyl acetylcitrate [FLno: 09.11] LD 0 values larger than 30 g/kg are reported (Annex IV, Table IV.1). 8.. Subacute, Subchronic, Chronic and Carcinogenicity Studies Subacute / subchronic / chronic toxicity data are available for five candidate substances, butoxyethanol [FLno: 0.4], butane1,3diol [FLno: 0.13], malonic acid [FLno: 08.03], glutaraldehyde [FLno: 0.149], nonanedioic acid [FLno: 08.3] and for supporting substances 1

16 Flavouring Evaluation Group Aliphatic primary and secondary saturated and unsaturated alcohols, aldehydes, acetals, carboxylic acids and esters containing an additional oxygenated functional group and lactones from chemical groups 9, 13 and 30 The EFSA Journal (00) 46 of the present Flavouring Group Evaluation (JECFA, 1998a; JECFA, 000c). (Annex IV, Table IV.) Available data on repeated dose toxicity show that haemolysis is the primary and critical response elicited in the main animal test models (rats and mice) following inhalation, oral or dermal exposure to butoxyethan1ol; where the haematotoxic action is produced by the metabolite butoxyacetic acid. Notably, the haematotoxic effect exhibits a pronounced species difference. In sensitive species (rat, mouse, hamster, baboon), butoxyethan1ol produces a characteristic toxicity that is revealed clinically by the appearance of haemoglobinuria and pathologically by changes in a variety of blood parameters (EURAR, 004a; EPA, 1999). LAELs of 69 and 8 mg/kg bw/day (slight decrease in body weight gain, haematological and liver effects) have been reported for male and female rats, respectively (NTP, 1993a). Human erythrocytes are about 0times less sensitive than rat erythrocytes as judged by prehaemolytic changes in vitro (increase in mean erythrocyte volume, erythrocyte deformability) consistently observed in both species. Studies have also shown that potentially sensitive human subpopulations, including children, the elderly and those with sickle cell anemia do not show increased sensitivity to the haemolytic action of butoxyethan1ol. Furthermore, the in vivo blood concentrations producing haemolysis in the animal experiments are considered unlikely to occur under normal conditions of human exposure to butoxyethan1ol (EURAR, 004a). The panel is aware that the LAEL for haematological effects found in the NTPstudy was used to derive an oral reference dose (RfD) of 0. mg/kg/day by the means of validated pharmacokinetic modelling (EPA, 1999). Step B4, however, asks for a NAEL and therefore, the RfD based upon a LAEL was not used for the present FGE. Carcinogenicity: In a year inhalation study, F344/N rats were exposed to 0, 31, 6. and 1 ppm and B6C3F1 mice were exposed to 0, 6., 1 and 0 ppm butoxyethan1ol; (NTP, 1993a), The exposure caused a low incidence of haemangiosarcoma in male mice at the highest exposure concentration; haemangiosarcoma did not occur in female mice or in rats. In female mice, butoxyethan1ol caused an increased incidence of forestomach tumours. It was not carcinogenic in rats. The occurrence of haemangiosarcoma in male mice only at highest exposure concentration is suggestive of a threshold phenomenon, related to the induction of haemolysis in rodent species. With regard to human relevance, the mechanism proposed for the induction of haemangiosarcomas strongly supports the conclusion that butoxyethan1ol is unlikely to be a carcinogenic hazard at the estimated level of intake as flavouring substance, because human erythrocytes are demonstrably more resistant to haemolysis than are rodent erythrocytes. Glutaraldehyde 4 [FLno: 0.149] (0, 0, 00 ppm in drinking water, resulting in doses of:.96.9, and mg/kg/day, respectively) was not tumorigenic in a year carcinogenicity study on male and female rats (Van Miller et al., 00). Furthermore, malonic acid [FLno: 08.03] was negative in a liver foci tumour promotion assay (Annex IV, Table IV.). 4 Glutaraldehyde is also used in food contact material (FCM). It was evaluated by the former Scientific Committee on Food (SCF List 7, however, this is not a final evaluation. According to German recommendations glutardialdehyde (synonym: glutaraldehyde) may be used for the production of artificial sausage skin (maximum use level 0.1 %). The maximum residual amount of glutardialdehyde is 0 mg per kg artificial sausage skin (ready for use). Furthermore, glutardialdehyde may be used as anti slime agent for the production of paper as FCM (maximum use level. % based on dry fibre material). The maximum residual amount of glutardialdehyde is mg per kg paper (ready for use). The Panel noted that maximum residual amounts of glutaraldehyde in food contact material (as set e.g. in German recommendations) could apparently conflict with reported use levels of glutaraldehyde as flavouring. However, in the German recommendations the maximum residual amounts were set considering the technologically needed use levels (limited data submitted) rather than on toxicological data, and the Panel therefore did not find the low maximum residual amounts for glutaraldehyde as such in conflict with higher use levels for glutaraldehyde as flavouring, which could therefore go through the Procedure. 16