The Japan Flavour and Fragrance Materials Association’s (JFFMA) safety assessment of acetal food flavouring substances uniquely used in Japan

Using the procedure devised by the Joint FAO/WHO Expert Committee on Food Additives (JECFA), we performed safety evaluations on five acetal flavouring substances uniquely used in Japan: acetaldehyde 2,3-butanediol acetal, acetoin dimethyl acetal, hexanal dibutyl acetal, hexanal glyceryl acetal and 4-methyl-2-pentanone propyleneglycol acetal. As no genotoxicity study data were available in the literature, all five substances had no chemical structural alerts predicting genotoxicity. Using Cramer’s classification, acetoin dimethyl acetal and hexanal dibutyl acetal were categorised as class I, and acetaldehyde 2,3-butanediol acetal, hexanal glyceryl acetal and 4-methyl-2-pentanone propyleneglycol acetal as class III. The estimated daily intakes for all five substances were within the range of 1.45–6.53 µg/person/day using the method of maximised survey-derived intake based on the annual production data in Japan from 2001, 2005, 2008 and 2010, and 156–720 µg/person/day using the single-portion exposure technique (SPET), based on the average use levels in standard portion sizes of flavoured foods. The daily intakes of the two class I substances were below the threshold of toxicological concern (TTC) – 1800 μg/person/day. The daily intakes of the three class III substances exceeded the TTC (90 μg/person/day). Two of these, acetaldehyde 2,3-butanediol acetal and hexanal glyceryl acetal, were expected to be metabolised into endogenous products after ingestion. For 4-methyl-2-pentanone propyleneglycol acetal, one of its metabolites was not expected to be metabolised into endogenous products. However, its daily intake level, based on the estimated intake calculated by the SPET method, was about 1/15 000th of the no observed effect level. It was thus concluded that all five substances raised no safety concerns when used for flavouring foods at the currently estimated intake levels. While no information on in vitro and in vivo toxicity for all five substances was available, their metabolites were judged as raising no safety concerns at the current levels of intake. Graphical Abstract


Flavouring substances in Japan
A wide variety of flavouring substances have been developed to improve the taste and aroma of foodstuffs but which have no nutritional properties. This wide variety enables the food industry to meet its needs to mimic the flavours of foods. According to the annual usage survey by the Japan Flavour and Fragrance Materials Association (JFFMA), approximately 3230 flavouring substances have been used commercially in Japan (Someya 2012). Among them, 129 substances are currently approved as designated additives by the Ministry of Health, Labour and Welfare, Japan. The remaining substances have been classified into 18 chemical structural groups and approved for use in Japan under the compliance of the Food Sanitation Act without a safety evaluation.

Safety evaluation procedures
In 1996, JECFA adopted a novel safety evaluation procedure for flavouring substances (WHO 1996). Using this procedure, the safety of many flavouring substances can be efficiently evaluated using the safety evaluation data for substances that have similar chemical structures or metabolic pathways, even if the substances under investigation have no safety data. Further evaluation protocols similar to the JECFA procedure are accepted for use in the United States and the European Union. In addition, the results of safety evaluations by JECFA, the Flavour and Extract Manufacturers Association (FEMA) Expert Panel, and EFSA are now widely used in more than 70 countries (Someya 2012;Konishi et al. 2014). Since 2003, the Food Safety Commission of Japan (FSCJ) has been performing safety evaluations on individually selected flavouring substances. Its conclusions are consistent with those obtained using the group evaluation by JECFA and, as of May 2014, 51 substances have already been designated as food additives in Japan based on these results (Someya 2012).
Therefore, of approximately 3230 flavouring substances used in Japan, about 2400 have been judged to be safe under limited conditions of use for flavouring by JECFA, FEMA, EFSA and FSCJ. However, this leaves about 800 flavouring substances whose safety has not been evaluated by any governmental or international organisation. Thus, in 2009, JFFMA established a Safety Reevaluation Committee, comprised of expert scientists in the fields of toxicology and food chemistry as well as representative members of JFFMA, to evaluate substances uniquely used in Japan using the JECFA evaluation procedure (see Figure S1 in the Supplemental data online), with the order of priority being determined by the annual volume of production.
Using the JECFA evaluation procedure, the present study evaluates the safety of five acetal flavouring substances uniquely used in Japan: acetaldehyde 2,3-butanediol acetal, acetoin dimethyl acetal, hexanal dibutyl acetal, hexanal glyceryl acetal and 4-methyl-2-pentanone propyleneglycol acetal. These substances have unique odour characteristics and are used with other flavouring substances to improve the taste of processed foods such as candy, chewing gum and soft drinks.

Prediction of genotoxic potential by chemical structure and in silico models
We evaluated the genotoxic potential of each acetal by analysing the biological potential of its chemical structure. While the JECFA decision tree approaches do not cover genotoxicity in its evaluation procedure (WHO 1996), the potential of genotoxicity is now evaluated before the decision tree steps by incorporating structural alerts and other available genotoxicity information from structurally related chemicals in JECFA (WHO 1998;1999a, 1999b) and FSCJ  (2005a, 2005b). In the present study, following the JECFA procedures, we examined structural alerts to chemical substructure groups, as is generally done for chemicals lacking genotoxicity information (Ashby & Tennant 1988Tennant et al. 1990). While genotoxicity study data were not available for any of the five substances, there were no structural alerts on their chemical structures. We also examined the genotoxic potential of the five substances using structureactivity relationship (SAR) analyses (Hansch 1969) using in silico prediction software that has been applied to the regulatory evaluation of flavouring substances (Cronin et al. 2003;Ono et al. 2012). We used three SAR software programs with different algorithms for analysis: DEREK (version 10.0.2; Lhasa Ltd, Leeds, UK), MultiCASE (version 1.90; Multicase Inc., Cleveland, OH, USA), and ADMEWORKS (version 4.0; Fujitsu Kyusyu Systems Ltd, Fukuoka, Japan). All the software programs used in the study calculated that no substances except acetaldehyde 2,3-butanediol acetal had the potential for genotoxicity. Regarding acetaldehyde 2,3-butanediol acetal, MultiCASE analysis gave a positive result for genotoxicity, but the other two programs gave negative results. Furthermore, alcohols and aldehydes or ketones, which are metabolites of acetals, have no genotoxic potential except for acetaldehyde (WHO 1999a(WHO , 1999b. In addition, the genotoxicity information of structurally related substances was used for evaluation. Based on these results, we decided that the JECFA evaluation procedure could be applied to these five acetal flavouring substances.
Estimating the daily intake of flavouring substances by the MSDI and SPET methods According to the annual production data from 2001,2005,2008 and 2010 in Japan, the total annual estimated volumes of acetaldehyde 2,3-butanediol acetal, acetoin dimethyl acetal, hexanal dibutyl acetal, hexanal glyceryl acetal and 4-methyl-2-pentanone propyleneglycol acetal were 2.02-5.51, 0.10-23.00, 5.85-11.97, 3.78-6.72 and 0-24.80 kg, respectively (JFFMA 2003(JFFMA , 2006(JFFMA , 2011unpublished information, 2008). Based on these data, the daily intake of each substance was calculated using the maximised surveyderived intake (MSDI; Young et al. 2006) and single-portion exposure technique (SPET; WHO 2008) procedures. The development of the MSDI procedure was based on disappearance data from periodic surveys of ingredient manufacturers using the volume of ingredients produced during the survey year (Young et al. 2006). Because the production volume changes every year, the maximum production volume of each substance was selected from the annual production data obtained to avoid any underestimation of the intakes. The resultant daily intakes per person of each of the five flavouring substances were estimated to be between 1.45 and 6.53 µg (Table 2).
Because the daily intakes estimated by the SPET method were higher than those by the MSDI method, it was decided to conduct a safety evaluation using the values calculated by the SPET method.

Estimation of absorption, metabolism and excretion
There was no information on the five acetals regarding absorption, metabolism and excretion from the JECFA assessment reports. Therefore, the components or structurally related substances of these five acetals were examined to estimate their metabolic fate. Among these structurally related acetals, 1,2,3-tris[(1´-ethoxy) ethoxy] propane was easily hydrolysed into acetaldehyde and glycerol (De Simone 1976). Potassium 2-(1 -ethoxy) ethoxypropanoate, an acetal of lactic acid, was completely hydrolysed into lactic acid, acetaldehyde and ethanol by incubation in gastric fluids (Moreno et al. 1984). 1,1-Dimethoxyethane, acetal and related acetals were hydrolysed within 1-5 h in simulated gastric fluid and, to a lesser extent, in simulated intestinal fluid (Morgareidge 1962). An ingested suspension of 1,1dimethoxyethane, acetal and other aliphatic acetals was rapidly hydrolysed within the stomach (Knoefel 1934). In a study examining the synthesis of valpronic acid (2propylpentanoic acid) by incubation of dimethyl, diethyl and diisopropyl acetals of 2-propylpentanal with an S10 fraction of rat liver homogenates or microsomal fraction, synthesis was inhibited by SKF-525A, a cytochrome P450 (CYP) inhibitor (Vicchio & Callery 1989). Synthesis was also inhibited by removal of nicotine adenine amide dinucleotide phosphoric hydrase or oxygen. These results were in accordance with the responses mediated by CYP. 2-Propyl-1-pentanol was detected as a major metabolite after incubation with the microsomal fraction after removing enzymes from the soluble fraction. Moreover, the incubation of deuterated diethyl acetal of 2-propylpentanal indicated the oxidation of ether methylene as a metabolism pathway of acyclic acetal. These results showed that acetalrelated substances were hydrolysed into aldehydes and alcohols.
Linear saturated acyclic alcohols, aldehydes and carboxylic acids, metabolites of acetals, are absorbed from the gastrointestinal tract (Dawson & Hullin 1954;von Oettingen 1960;Brabec 1993;Katz & Guest 1994;Lington & Bevan 1994). While most of these are endogenous biogenic substances making it difficult for their plasma half-life to be measured (Lington & Bevan 1994), these acetal metabolites are known to be further metabolised through β-oxidation and tricarboxylic acid pathways into innocuous products.
Aliphatic ketones are excreted into urine as glucuronic acid conjugates through reduction to secondary alcohols (Neubreuer 1901;Leibman 1971;Bosron & Li 1980;Felsted & Bachur 1980). In the case of short-chain aliphatic ketones, it is possible that they are excreted directly into expired air or undergo ω-oxidation or ω-1-oxidation (Haggard et al. 1945;Scopinaro et al. 1947;Saito 1975;Brown et al. 1987;Schwartz 1989), metabolised into the corresponding carboxylic acids, and then metabolised into innocuous substances through the tricarboxylic acid pathway. Although the metabolism of short-chain aliphatic ketones through ω-oxidation and ω-1-oxidation are competitive with long-chain aliphatic ketones when they exist at high concentrations (Dietz et al. 1981;Topping et al. 1994), such competition does not occur at the intake levels when used as a flavouring substance.
Orally administered acetoin is rapidly absorbed from the gastrointestinal tract in rats (Gabriel et al. 1972). In humans, aliphatic acyclic α-diketones are metabolised into ketocarboxylic acids through α-hydroxylation and oxidation of the terminal methyl group, when they exist at low concentrations (WHO 1999b). α-Ketocarboxylic acids are metabolised into simple aliphatic carboxylic acids and carbon dioxide through oxidative decarboxylation and further into innocuous products through β-oxidation and the tricarboxylic acid cycle (WHO 1999b). In rats, rabbits and dogs, aliphatic acyclic α-diketones are reduced to diols and further conjugated with glucuronic acid to excrete into urine when they are exposed to animals at high concentrations (Westerfeld & Berg 1943;Gabriel et al. 1972;Otsuka et al. 1996).
Safety evaluation of the five substances according to the JECFA evaluation procedure Decision tree approach Details of the JECFA evaluation procedure are described in WHO (1996). A summary of the decision tree approach for this procedure is shown in Figures S1 and S2 in the Supplemental data online.
Step 1: In the first step, each flavouring substance was subjected to classification into three chemical structure classes: I, II, or III, according to Cramer's decision tree, as shown in Figure S2 (Cramer et al. 1976). Substances in class I have a simple chemical structure with known efficient metabolic pathways that would suggest a low oral toxicity. Substances in class II are less innocuous than class I, but do not contain structural features suggestive of having toxicity like substances in class III. Substances in class III have a chemical structure that may cause significant toxicity.
According to the results of the Cramer decision tree ( Figure S2; Cramer et al. 1976), two substances, acetoin dimethyl acetal and hexanal dibutyl acetal, were assigned to structural class I, and the remaining three substances, acetaldehyde 2,3-butanediol acetal, hexanal glyceryl acetal and 4-methyl-2-pentanone propyleneglycol acetal, to structural class III (Table 3). In addition, hydrolysed products of aliphatic acetals, i.e., alcohols and aldehydes or ketones, which are components of these substances, also belong to structural class I.
Step 2: In the second step, the metabolic fate of each substance was predicted by chemical structure and applied using a decision-tree approach consisting of two branches: one for substances that are metabolised into innocuous products, and the other for substances that are not. This second step showed that all five substances were easily hydrolysed into alcohols and aldehydes or ketones by carboxylesterases or esterases, and the resultant substances are metabolised into innocuous products as mentioned above (WHO 1999b) (Table 4). Then the evaluation of all substances proceeded to the A-side of the decision tree ( Figure S1, Step A3).
Step A3: In this step, the estimated daily intake of each substance was compared with the threshold level of human exposure, according to Cramer's structural classes (Cramer et al. 1976), in which the estimated threshold levels are 1800, 540, and 90 µg/person/day for structural classes I, II, and III, respectively (Munro et al. 1996(Munro et al. , 1999. As mentioned above, the estimated daily intakes of the five acetal flavouring substances as calculated by the SPET method were higher than those by the MSDI method. The daily intake of two class I substances were within the range of 156-200 μg kg -1 day -1 (Table 4). These intake values were below the threshold of toxicological concern (TTC) applied to structural class I substances. However, the daily intake level of three class III substances were within the range of 300-720 μg kg -1 day -1 , values greater than the TTC. Therefore, the evaluation of these three substances proceeds to step A4.
Step A4: None of the three structural class III substances is a normal constituent of the body. It is expected that acetaldehyde 2,3-butanediol acetal is hydrolysed into acetaldehyde and 2,3-butanediol, and hexanal glyceryl acetal into hexanal and glycerol. These metabolites are normal constituents of the body (Table 4). On the other hand, it is expected that 4-methyl-2-pentanone propyleneglycol acetal is hydrolysed into 4-methyl-2-pentanone and propyleneglycol. Propyleneglycol is expected to be further metabolised into endogenous substances; however, 4methyl-2-pentanone is not. Therefore, the evaluation of this substance proceeds to step A5.
Step A5: A no-observed-effect-level (NOEL) of 50 mg kg -1 day -1 was reported in a 13-week study using Sprague-Dawley rats for 4-methyl-2-pentanone (Microbiological Associates Inc. 1986), a metabolite of 4-methyl-2-pentanone propyleneglycol acetal that is not a normal constituent of the body. This NOEL provides a safety margin of > 15 000 for the intake of 4-methyl-2pentanone calculated by the daily intake level of 4-  Table 4. Safety evaluation of acetal flavouring substances according to the JECFA evaluation procedure -Steps 2, A3, A4 and A5.

Consideration of the combined intakes from use as flavouring agents
Because JECFA considers that evaluating the combined intake is not necessary if the highest MSDI value is less than 20 μg day -1 (WHO 2011), we decided to omit the evaluation of combined intakes for all five substances.
Assessing toxicological information for hydrolysed products of the five flavouring substances As no information on in vitro and in vivo toxicity for all five substances was available, instead their metabolites were examined using the data tabulated by JECFA (WHO 1998;1999a, 1999bFSCJ 2005aFSCJ , 2005b.

Subacute toxicity
The subacute toxicity by oral administration of acetaldehyde, acetoin, hexanal, butyl alcohol and methyl isobutyl ketone (MIBK; also known as 4-methyl 2-pentanone) in rats has been reported in several studies (Table 5). There are no subacute toxicity data by oral administration of 2,3butanediol, methanol, glycerol and propylene glycol. In a study of Wistar rats (24 per group) given drinking water containing acetaldehyde for 11 weeks, liver cell fatty degeneration was detected at 500 mg kg -1 day -1 (Matysiak-Budnik et al. 1996). In a study of Wistar rats (10 per group) given drinking water containing acetaldehyde for 4 weeks, an increase in relative kidney weight and thickening of the limiting ridge of the forestomach were noted at 625 mg kg -1 day -1 (Til et al. 1986).
In a study of CFE rats (15 per group) given drinking water containing acetoin for 13 weeks, the body weight gain was reduced at 12 000 mg kg −1 (1300 mg kg -1 day -1 ) (Gaunt et al. 1972). Low values in erythrocyte counts and haemoglobin concentration were also noted at 1300 mg kg -1 day -1 . However, haematocrit values were not influenced by acetoin treatment. There were also no toxicologically relevant effects on the clinical signs, urinalysis, blood chemistry or histopathological examination.
Another study of Sprague-Dawley rats (10 per group) given drinking water containing hexanal for 4 weeks, there were no toxicologically relevant effects in the treated groups (Komsta et al. 1988).
A study of Sprague-Dawley rats (30 per group) orally administered butyl alcohol for 13 weeks showed increased incidences of ataxia or hypokinesia in both sexes at 500 mg kg -1 day -1 from weeks 7 to 13 (USEPA 1987).
In a study of Sprague-Dawley rats (30 rats per group) orally administered MIBK for 13 weeks, high values for haemoglobin concentration and haematocrit and low values for lymphocytes at 1000 mg kg -1 day -1 were observed. Blood biochemically, a high level of serum alanine aminotransferase was observed at ≥ 250 mg kg -1 day -1 . At 1000 mg kg -1 day -1 , there were high levels of serum alkaline phosphatase and total protein, high levels for total cholesterol, and low values for the albumin/globulin ratio, high values for blood urea nitrogen and serum potassium and low values for serum glucose. In addition, increases in kidney and liver weights were observed at doses of ≥ 250 and of 1000 mg kg -1 day -1 , respectively. Increases in urinary protein and ketones were observed at 1000 mg kg -1 day -1 . Histopathologically, an increased incidence of mild nephropathy was observed at 1000 mg kg -1 day -1 (Microbiological Associates Inc. 1986). In a study of a group of five HLA, Wistar rats given drinking water containing MIBK for 120 days revealed increases in kidney weights at 1040 mg kg -1 day -1 (Carnegie-Mellon Institute of Research 1977a, 1977b).

Long-term toxicity/carcinogenicity
No information was available on chronic toxicity of the components or structurally related substances of these acetals. Regarding carcinogenicity, there are several sets of data on acetaldehyde. Groups of Sprague-Dawley rats were given drinking water containing acetaldehyde for 104 weeks (Soffritti et al. 2002). The total incidence of malignant tumours increased in both sexes at a concentration of 2500 mg l -1 . However, in their evaluation report of acetaldehyde, FSCJ determined that the study results by Soffritti et al. (2002) could only be used as reference data, because the tumours were sporadically observed without showing dose-related increases or any apparent target organs (FSCJ 2005a).
In an inhalation carcinogenicity study of acetaldehyde targeting the nasal mucosa in rats, cancer developed at doses of 750, 1500 and 3000 mg kg −1 (the dose was reduced from 3000 to 1000 mg kg −1 after 11 months), 6 h day -1 , 5 days week -1 for the maximum period of 27 months. It also developed when targeting the larynx in hamsters at a dose of 2500 mg kg −1 (the dose was later reduced to 1650 mg kg −1 ), 7 h day -1 , 5 days week -1 for 52 weeks (IARC 1999). However, FSCJ (2005a) has judged that these inhalation study data are irrelevant for oral exposure to food additives.

Genotoxicity
Several in vivo and in vitro genotoxicity studies of acetaldehyde, acetoin, hexanal, butyl alcohol and 4-methyl 2pentanone, which are components of acetals, have been reported (Obe & Ristow 1977;Florin et al. 1980;Garst et al. 1983;He & Lambert 1985;Marnett et al. 1985; Soffritti et al. (2002) and FSCJ (2005a) Acetoin CFE rat (male/female) Drinking water Carnegie-Mellon Institute of Research (1977aResearch ( , 1977b Note: a There were no short-term/subchronic toxicological reports of 2,3-butanediol, methanol, glycerol and propylene glycol. Norppa et al. 1985;Kim et al. 1987;Wangenheim & Bolcsfoldi 1988;Brooks et al. 1988;O'Donoghue et al. 1988;Brambilla et al. 1989;Furnus et al. 1990;Helander & Lindahl-Kiessling 1991;Dillon et al. 1992;Zeiger et al. 1992;Kapp et al. 1993;Müller et al. 1993;Martelli et al. 1994;WHO 1998;FSCJ 2005b). The results of these studies are summarised in Table 6. The Ames test for acetaldehyde showed negative results (Dillon et al. 1992), but mutagenicity in a mouse lymphoma thymidine kinase test with and without metabolic activation revealed positive results (Wangenheim & Bolcsfoldi 1988). The chromosomal aberration test for acetaldehyde using human lymphocytes showed negative results, but positive results using lymphocytes from a patient suffering from Fanconi anaemia (Obe et al. 1979). Acetaldehyde increased the frequency of DNA strand breakage in mice by inhalation exposure (IARC 1999), as well as that of sister chromatid exchange in cultured human lymphocytes (He & Lambert 1985;Norppa et al. 1985) and in mice by intraperitoneal administration (IARC 1999). Acetaldehyde also increased the frequency of micronucleus formation in mice by intraperitoneal administration (FSCJ 2005a). On the other hand, negative results were detected in micronucleus tests using germ cells by intraperitoneal administration (IARC 1999). DNA adduct formation in humans was detected in granulocytes and lymphocytes in the blood caused by acetaldehyde through alcohol intake (Fang & Vaca 1997;Brooks & Zakhari 2014).
Positive results were reported in in vitro forward-or reverse-mutation assays of acetoin, hexanal and butyl alcohol (Garst et al. 1983;Brambilla et al. 1989;WHO 1998), but negative results were detected in other mutation tests. 4-Methyl 2-pentanone showed no genotoxicity in either in vitro or in vivo assays (WHO 1999a(WHO , 1999b. While acetaldehyde is considered to be genotoxic, JECFA concluded that it does not raise any safety concerns at the current levels of intake because acetaldehyde can be predicted to undergo complete metabolism to endogenous products via the fatty acid and tricarboxylic acid pathways (WHO 1998). With regard to acetoin, hexanal and butyl alcohol that showed some positive results in genotoxicity studies, these hydrolysed products belong to structural class I, according to the results of the Cramer decision tree. Therefore, it does not raise any safety concerns at the current levels of intake because hydrolysed products can be predicted to undergo complete metabolism to endogenous products.
Evaluation of acetals based on JECFA's procedure This review has presented the key scientific data relevant to the safety evaluation of acetals for use as food flavours using the JECFA evaluation procedure. Regarding the prediction of genotoxicity, four of the five substances had no structural alerts so it was considered that these substances had no genotoxic potential based on in silico software analyses. While genotoxicity study data are not available for all five acetals of concern, it is considered that these substances can be rapidly hydrolysed into alcohols and aldehydes or ketones. Moreover, structurally acetal-related substances are also reported to be hydrolysed into aldehydes and alcohols. In contrast, according to the Cramer's decision tree (Cramer et al. 1976), two of the five substances, acetoin dimethyl acetal and hexanal dibutyl acetal, were categorised as class I and the other three substances, acetaldehyde 2,3-butanediol acetal, hexanal glyceryl acetal and 4-methyl-2-pentanone propyleneglycol acetal, were categorised as class III. The estimated daily intakes for each of these substances were between 1.45 and 6.53 µg/person/day or between 156 and 720 µg/person/day, using the MSDI or SPET method, respectively. The daily intake of the two class I substances were below the toxicological threshold of concern of 1800 μg/person/day. Conversely, the daily intake of the three class III substances exceeded the TTC of 90 μg/person/day. Two of them, acetaldehyde 2,3-butanediol acetal and hexanal glyceryl acetal, were immediately metabolised after ingestion and excreted as endogenous products. Ingested 4-methyl-2-pentanone propyleneglycol acetal is hydrolysed into 4-methyl-2pentanone and propyleneglycol. Among them, propyleneglycol will be metabolised into endogenous products (WHO 2002). However, 4-methyl-2-pentanone is not endogenously metabolised with a NOEL of 50 mg kg -1 day -1 , which is more than 15 000 times the daily intake level of 4-methyl-2-pentanone propyleneglycol acetal determined by the SPET method. Thus, it was concluded that all five substances raise no safety concerns when used for flavouring foods at the currently estimated intake levels.

Evaluation of toxicity data of acetal metabolites
The subacute toxicity studies of acetal metabolites revealed no major toxicity with these compounds. Available studies on the genotoxicity of acetal components also provided negative results except for some positive results with acetaldehyde. Acetaldehyde was also reported to be carcinogenic by high-dose inhalation exposure in rodents.
JECFA has previously concluded that acetaldehyde does not raise any safety concerns at the current levels of intake because ingested acetaldehyde will undergo complete metabolism to endogenous products and its level do not exceed the physiological range (WHO 1998). On the other hand, there is genetic-epidemiological evidence showing that humans deficient in the oxidation of acetaldehyde to acetate have a substantially increased Table 6. Summary of in vitro and in vivo genotoxicity data for acetaldehyde, butanol, hexanal, acetoin and 4-methyl-2-pentanone.  Garst et al. (1983) Reverse mutation Salmonella typhimurium (TA100) 420 μg/plate Negative a Kim et al. (1987) Hexanal Forward mutation assay V79 Chinese hamster lung cells 3-30 mmol l -1 Positive a Brambilla et al. (1989) Unscheduled DNA synthesis assay Adult human and rat hepatocytes 3-100 mmol l -1 Negative a Martelli et al. (1994) Ames test Salmonella typhimurium (TA102, TA104) Up to 1 mg/plate Negative a Marnett et al. (1985) Ames test (spot test) Salmonella typhimurium (TA98, TA100, TA1535, TA1537) 3 µmol/plate Negative b Florin et al. (1980) Forward mutation assay V79 Chinese hamster lung cells risk for development of alcohol-related cancers, in particular of the oesophagus and the upper aero-digestive tract (Eriksson 2015). It has been reported that there is a relationship between the genetic polymorphism of alcohol dehydrogenase (ALDH) and alcohol metabolism. It is known that ALDH II-type deficiency is more common among Orientals (Yoshida et al. 1984). Although ALDH II-type deficiency is likely to increase the blood aldehyde levels in more susceptible humans than in less susceptible ones (Enomoto et al. 1991), other metabolic pathways are considered to function in a complementary manner (Kunitoh et al. 1996;Riveros-Rosas et al. 1997). It is reported that approximately 1.3-3.9 μM of blood acetaldehyde levels can be detected in healthy persons (Lynch et al. 1983;Fukunaga et al. 1993). In Japan, the blood acetaldehyde level in man is not considered to exceed 14 μM even if the estimated daily intake of acetaldehyde (approximately 19 mg/person/day) is consumed at a time, and if 100% of it is absorbed and distributed in the body without being metabolised by the first-pass effect, as reported by FSCJ (2005a). However, the situation, in which the level of acetaldehyde, equivalent to that used as a flavouring ingredient, is consumed in daily dietary life, is far from being applicable to the above hypothesis. In fact, not all the orally ingested acetaldehyde is absorbed directly into the body, the majority of it is supposed to change into acetic acid via metabolism by ALDH and other enzymes in the gastrointestinal tract and liver (Tsutsumi et al. 1988;Yin et al. 1994;Kunitoh et al. 1996;Riveros-Rosas et al. 1997). JECFA has also concluded that acetaldehyde does not raise any safety concerns at the current levels of intake because ingested acetaldehyde will undergo complete metabolism to endogenous products.

Conclusions
In conclusion, the five acetals, acetaldehyde 2,3-butanediol acetal, acetoin dimethyl acetal, hexanal dibutyl acetal, hexanal glyceryl acetal and 4-methyl-2-pentanone propyleneglycol acetal, for flavouring foods pose no health risk to humans, and the intake of each substance as a foodflavouring ingredient is safe at the present levels of use.

Disclosure statement
Hiroyuki Okamura, Kenji Saito, Fumiko Sekiya, Shim-Mo Hayashi, Yoshiharu Mirokuji and Shinpei Maruyama are employed by flavour manufacturers whose product lines include flavouring substances. The views and opinions expressed in this article are those of the authors and not necessarily those of their respective employers. Hajime Abe, Yasuko Hasegawa-Baba, Atsushi Ono, Madoka Nakajima, Masakuni Degawa, Shogo Ozawa, Makoto Shibutani and Tamio Maitani declare that no conflicts of interest exist.

Funding
This work was supported by the Japan Flavour and Fragrance Materials Association.