The Committee evaluated a group of flavouring agents¹ comprising 46 structurally related substances using the Procedure for the Safety Evaluation of Flavouring Agents (see Figure 1, Introduction). All members of this group are aromatic primary alcohols, aldehydes, carboxylic acids, or their corresponding esters or acetals. The structural feature common to all members of the group is a primary oxygenated functional group bound directly to a benzene ring. The ring also contains hydroxy or alkoxy substituents (see Table 1).

Flavouring agent	No.	CAS No. and structure	Step A3b Does intake exceed the threshold for human intake?	Step A4 Is the flavouring agent or are its metabolites endogenous?	Step A5 Adequate margin of safety for the flavouring agent or related substance?	Comments on predicted metabolism	Conclusion based on current intake
Structural class I
4-Hydroxybenzyl alcohol	955	623-05-2	No Europe: 6 USA: 0.06	NR	NR	See note 1.	No safety concern
4-Hydroxybenzaldehyde	956	123-08-0	No Europe: 64 USA: 59	NR	NR	See note 1.	No safety concern
4-Hydroxybenzoic acid	957	99-96-7	No Europe: 19 USA: 17	NR	NR	See note 1.	No safety concern
2-Hydroxybenzoic acid	958	69-72-7	No Europe: 0.03 USA: 0.03	NR	NR	See note 1.	No safety concern
Butyl para-hydroxy benzoatec	870	94-26-8	No Europe: ND USA: 0.03	NR	NR	See note 2.	Evaluation not finalized
Anisyl alcohol	871	105-13-5	No Europe: 150 USA: 58	NR	NR	See note 1.	No safety concern
Anisyl formate	872	122-91-8	No Europe: 46 USA: 24	NR	NR	See note 2.	No safety concern
Anisyl acetate	873	104-21-2	No Europe: 59 USA: 300	NR	NR	See note 2.	No safety concern
Anisyl propionate	874	7549-33-9	No Europe: ND USA: 5	NR	NR	See note 2.	No safety concern
Anisyl butyrate	875	6963-56-0	No Europe: 34 USA: 0.1	NR	NR	See note 2.	No safety concern
Anisyl phenylacetate	876	102-17-0	No Europe: 0.003 USA: 0.1	NR	NR	See note 4.	No safety concern
Veratraldehyde	877	120-14-9	No Europe: 140 USA: 55	NR	NR	See note 1.	No safety concern
para-Methoxybenzaldehyde	878	123-11-5	No Europe: 440 USA: 580	NR	NR	See note 1.	No safety concern
para-Ethoxybenzaldehyde	879	10031-82-0	No Europe: 0.1 USA: 0.01	NR	NR	See note 1.	No safety concern
Methyl ortho-methoxybenzoate	880	606-45-1	No Europe: 57 USA: 8	NR	NR	See note 2.	No safety concern
2-Methoxybenzoic acid	881	579-75-9	No Europe: ND USA: 0.01	NR	NR	See note 1.	No safety concern
3-Methoxybenzoic acid	882	586-38-9	No Europe: ND USA: 0.01	NR	NR	See note 1.	No safety concern
4-Methoxybenzoic acid	883	100-09-4	No Europe: ND USA: 0.1	NR	NR	See note 1.	No safety concern
Methyl anisate	884	121-98-2	No Europe: 1 USA: 0.01	NR	NR	See note 2.	No safety concern
Ethyl para-anisate	885	94-30-4	No Europe: 11 USA: 2	NR	NR	See note 2.	No safety concern
Vanillyl alcohol	886	498-00-0	No Europe: 6 USA: 6	NR	NR	See note 1.	No safety concern
Vanillind	889	121-33-5	Yes Europe: 55 000 USA: 150 000	No	Yes; the NOEL of 1000 mg/kg bw per day in a 2-year study in rats is > 100 times the estimated daily intake of vanillin when used as a flavouring agent.	See note 1.	No safety concern
4-Hydroxy-3-methoxy-benzoic acid	959	121-34-6	No Europe: 29 USA: 26	NR	NR	See note 1.	No safety concern
Vanillin acetate	890	881-68-5	No Europe: 2 USA: 1	NR	NR	See note 2.	No safety concern
Vanillin isobutyrate	891	20665-85-4	No Europe: 64 USA: 0.04	NR	NR	See note 2.	No safety concern
Salicylaldehyde	897	90-02-8	No Europe: 98 USA: 16	NR	NR	See note 1.	No safety concern
2-Hydroxy-4-methyl-benzaldehyde	898	698-27-1	No Europe: 1 USA: 0.3	NR	NR	See note 1.	No safety concern
Methyl salicylatee	899	119-36-8	Yes Europe: 490 USA: 44 000	No	Yes; the NOEL of 50 mg/kg bw per day in a 2-year study in dogs is > 100 times the estimated daily intake of methyl salicylate when used as a flavouring agent.	See note 2	No safety concern
Ethyl salicylate	900	118-61-6	No Europe: 31 USA: 1700	NR	NR	See note 2.	No safety concern
Butyl salicylate	901	2052-14-4	No Europe: 0.01 USA: 0.0007	NR	NR	See note 2.	No safety concern
Isobutyl salicylate	902	87-19-4	No Europe: 1 USA: 6	NR	NR	See note 2.	No safety concern
Isoamyl salicylate	903	87-20-7	No Europe: 49 USA: 7	NR	NR	See note 2.	No safety concern
Benzyl salicylate	904	118-58-1	No Europe: 30 USA: 29	NR	NR	See note 4.	No safety concern
Phenethyl salicylate	905	87-22-9	No Europe: 0.2 USA: 4	NR	NR	See note 4.	No safety concern
ortho-Tolyl salicylate	907	617-01-6	No Europe: ND USA: 30	NR	NR	See note 4.	No safety concern
2,4-Dihydroxybenzoic acid	908	89-86-1	No Europe: ND USA: 6	NR	NR	See note 1.	No safety concern
Structural class II
Vanillyl ethyl ether	887	13184-86-6	No Europe: 22 USA: 22	NR	NR	See note 1.	No safety concern
Vanillyl butyl ether	888	82654-98-6	No Europe: ND USA: 0.1	NR	NR	See note 1.	No safety concern
Ethyl vanillinf	893	121-32-4	Yes Europe: 6200 USA: 43 000	No	Yes; the NOEL of 500 mg/kg bw per day in a 14-week study in rats is > 100 times the estimated daily intake of ethyl vanillin when used as a flavouring agent.	See note 1.	No safety concern
Vanillin erythro & threo- butan-2,3-diol acetal	960		No Europe: 4 USA: 3	NR	NR	See note 2.	No safety concern
Ethyl vanillin isobutyrate	953	188417-26-7	No Europe: 64 USA: ND	NR	NR	See note 2.	No safety concern
Ethyl vanillin propylene glycol acetal	954	68527-76-4	No Europe: 39 USA: 36	NR	NR	See note 2.	No safety concern
Piperonyl acetate	894	326-61-4	No Europe: 41 USA: 11	NR	NR	See note 3.	No safety concern
Piperonyl isobutyrate	895	5461-08-5	No Europe: 0.1 USA: 3	NR	NR	See note 3.	No safety concern
Piperonalg	896	120-57-0	Yes Europe: 1700 USA: 3200	No	Yes; the NOEL of 250 mg/kg bw per day in a 2-year study in rats is > 100 times the estimated daily intake of piperonal when used as a flavouring agent	See note 3.	No safety concern
Ethyl vanillin beta-d-gluco-pyranoside	892	122397-96-0	No Europe: ND USA: 30	NR	NR	See note 2.	No safety concern

The Committee previously evaluated four members of this group. Ethyl vanillin (No. 893) was evaluated at by the Committee at its eleventh meeting (Annex 1, reference 14), when a conditional ADI of 0–10 mg/kg bw was allocated. At its thirty-fifth meeting, the Committee converted this value to a temporary ADI of 0–5 mg/kg bw (Annex 1, reference 88). At its thirty-ninth meeting, the Committee extended the temporary ADI (Annex 1, reference 101). At its forty-fourth meeting, the Committee allocated an ADI of 0–3 mg/kg bw to ethyl vanillin (Annex 1, reference 116). Vanillin (No. 889) was evaluated by the Committee at its eleventh meeting (Annex 1, reference 14), when an ADI of 0–10 mg/kg bw was established. Methyl salicylate (No. 899) was evaluated at the eleventh meeting (Annex 1, reference 14), when an ADI of 0–0.5 mg/kg bw was allocated. Piperonal (No. 896) was also evaluated at the eleventh meeting, when an ADI of 0–2.5 mg/kg bw was allocated (Annex 1, reference 14).

At its present meeting, the Committee evaluated a group of benzyl derivatives, none of which contains hydroxy- or alkoxy- substituents. Twenty-nine of the 46 members of this group of flavouring agents have been reported to occur naturally in food. Vanillin, a major constituent of natural vanilla, is also present in strawberries and milk. Methyl salicylate, the predominant substituent of oil of wintergreen, is also found in tomatoes and grilled beef. Ethyl vanillin has been detected in raspberries and ginger, while piperonal is found in cooked chicken and pepper (Maarse et al., 1999).

The total annual production of the 46 flavouring agents in this group is 450 000 kg in Europe (International Organization of the Flavor Industry, 1995) and 1 800 000 kg in the USA (Lucas et al., 1999). Vanillin (No. 889), ethyl vanillin (No. 893), methyl salicylate (No. 899), and piperonal (No. 896), for which ADIs were previously established by the Committee, accounted for approximately 98% of the total annual volume in Europe and 99% in the USA. In Europe, the estimated daily per capita intakes of these compounds were 55 mg of vanillin, 6.2 mg of ethyl vanillin, 0.5 mg of methyl salicylate, and 1.6 mg of piperonal. In the USA, the estimated daily per capita intakes were 150 mg of vanillin, 43 mg of ethyl vanillin, 44 mg of methyl salicylate, and 3.2 mg of piperonal. The estimated daily per capita intakes of two other flavouring agents in this group were in the low milligram range: the estimated intake of ethyl salicylate is 1.7 mg/person per day in the USA and that of para-methoxybenzaldehyde is around 0.5 mg/person per day in both Europe and the USA. The estimated intakes of most of the other 40 flavouring agents were between 1 and 100 µg/person per day, and those of 10 were < 1 µg/person per day. The annual usage volume of these substances in Europe and the USA is shown in Table 2.

The aromatic esters in this group are expected to be appreciably hydrolysed through the catalytic activity of the intestinal carboxylesterases, especially beta-esterases, to benzyl alcohol or benzoic acid derivatives before absorption (Heymann, 1980; Anders, 1989). Likewise, acetals of substituted benzaldehyde derivatives are hydrolysed in gastric and intestinal fluids to yield benzaldehyde and component aliphatic alcohols. The resulting hydroxy- and alkoxy-substituted benzyl derivatives are rapidly absorbed in the gut, metabolized in the liver, and excreted in urine (Davison, 1971; Abdo et al., 1985; Temellini et al., 1993).

Once absorbed, benzyl derivatives are oxidized to the corresponding benzoic acid derivative, which is subsequently excreted unchanged or as sulfate or glucuronide conjugates (Sammons & Williams, 1946; Davison, 1971; Scheline, 1972). Piperonal is oxidized to piperonylic acid and excreted mainly as the glycine conjugate (Kamienski & Casida, 1970; Klungsoyr & Scheline, 1984). Minor metabolic detoxication pathways include O-demethylation, reduction, and decarboxylation. These pathways arise during enterohepatic cycling of conjugated benzyl metabolites and subsequent intestinal bacterial action (Strand & Scheline, 1975).

Step 1 Thirty-six of the 46 flavouring agents in this group were assigned to structural class I (Cramer et al., 1978). They are simple substituted aromatic compounds that are expected to be hydrolysed to aromatic aldehydes and simple aliphatic alcohols (Nos 870–875, 877–886, 889–891, 897–903 and 955–959), or they contain two aromatic rings that are expected to be hydrolysed to mononuclear residues with simple functional groups (Nos 876, 904, 905, and 907). The remaining 10 flavouring agents (Nos 887, 888, 892–896, 953, 954, and 960) are ethyl vanillin or piperonal derivatives that contain either an ethoxy or a methylene dioxy substituent. They are common components of food or closely related to common components of food and were assigned to structural class II.

Step 2 At current levels of intake, none of the 46 flavouring agents would be expected to saturate the metabolic pathways, and all are predicted to be metabolized to innocuous products.

Step A3 The estimated per capita intakes of 42 of the 46 flavouring agents in this group were below the human intake threshold for their respective classes (i.e., 1800 µg/day for structural class I and 540 µg/day for structural class II).

These 42 flavouring agents therefore pose no safety concern when used at their current levels.

The estimated intakes of vanillin (No. 889) and methyl salicylate (No. 899), which are in class I, exceed the daily intake threshold of 1800 µg/person. The estimated intakes of vanillin are 55 000 µg/person per day in Europe and 150 000 µg/person per day in the USA, and that of methyl salicylate is 44 000 µg/person per day in the USA. The estimated intakes of ethyl vanillin (No. 893) and piperonal (No. 896), in class II, exceed the intake threshold of 540 µg/person per day. The estimated intakes of ethyl vanillin are 6200 µg/person per day in Europe and 43 000 µg/person per day in the USA. For piperonal, the estimated intakes are 1700 µg/person per day in Europe and 3200 µg/person per day in the USA) (International Organization of the Flavor Industry, 1995; Lucas et al., 1999).

The estimated intakes of these four substances are below their respective ADI values. The daily intakes of vanillin in Europe and the USA, approximately 0.9 and 2.5 mg/kg bw, do not exceed the ADI of 0–10 mg/kg bw for vanillin. The highest estimated daily intakes of ethyl vanillin (0.7 mg/kg bw in the USA) and piperonal (0.05 mg/kg bw in the USA) do not exceed the ADIs of 0–3 mg/kg bw for ethyl vanillin and 0–2.5 mg/kg bw for piperonal. The estimated daily intake of the remaining substance, methyl salicylate, is 0.7 mg/kg bw, which is approximately equal to its ADI of 0–0.5 mg/kg bw.

The estimates of intake based on total annual production include the assumption that only 10% of the population eats these agents. The Committee received a report on the intake of methyl salicylate based on individual dietary records and its use in the USA in baked goods, chewing-gum, hard and soft sweets, and beverages (Edelstein, 2=01). The analysis showed that > 50% of the population would be expected to eat methyl salicylate. Use of this measured proportion of eaters in place of the default assumption of 10% yields an estimated intake of methyl salicylate of 0.1 mg/kg bw, which is below the current ADI of 0–0.5 mg/kg bw.

Step A4 Vanillin (No. 889), methyl salicylate (No. 899), ethyl vanillin (No. 893), and piperonal (No. 896) are not endogenous in humans.

Step A5 The ADI of 0–10 mg/kg bw for vanillin is based on a NOEL of 1000 mg/kg bw per day in a 2-year feeding study in rats (Hagan et al., 1967). This NOEL provides a margin of safety that is about 400 times the per capita intake of vanillin from its current use as a flavouring agent in Europe (0.9 mg/kg bw per day) or in the USA (2.5 mg/kg bw per day).

The ADI of 0–0.5 mg/kg bw for methyl salicylate is based on a NOEL of 50 mg/kg bw per day reported in a 2-year study in dogs (Webb & Hansen, 1963). This NOEL is more than 1000 times greater than the intake of methyl salicylate from its current use as a flavouring agent in Europe (0.008 mg/kg bw per day). This NOEL is also more than 100 times greater than the intake of methyl salicylate in the USA when intake is calculated by using the measured portion of eaters of 50% (0.1 mg/kg bw per day).

A NOEL of 500 mg/kg bw per day of ethyl vanillin was reported in a 14-week feeding study in rats (Hooks et al., 1992). This NOEL is about 700 times greater than the intake of ethyl vanillin from its use as a flavouring agent in Europe (0.1 mg/kg bw per day) or in the USA (0.7 mg/kg bw per day).

A NOEL of 250 mg/kg bw per day for piperonal was reported in a 2-year study in rats (Bar & Griepentrog, 1967). This NOEL is more than 5000 times the intake of piperonal from its use as a flavouring agent in Europe (0.028 mg/kg bw per day) and in the USA (0.05 mg/kg bw per day).

The considerations on intake and other information used to perform the stepwise evaluations of the 46 hydroxy- and alkoxy-substituted benzyl derivatives according to the Procedure are summarized in Table 1.

In the unlikely event that all 36 flavouring agents in structural class I were consumed on a daily basis, the estimated combined intake would exceed the human intake threshold (1800 µg/person per day). The estimated combined intake of all class II substances would also exceed the intake threshold (540 µg/person per day). However, all 46 substances are expected to be efficiently detoxicated and would not saturate the available detoxication pathways. On the basis of the evaluation of the collective data, combined intake would present no safety concerns.

The Committee retained the previously established ADIs of 0–10 mg/kg bw for vanillin (No. 889), 0–3 mg/kg bw for ethyl vanillin (No. 893), 0–2.5 mg/kg bw for piperonal (No. 896), and 0–0.5 mg/kg bw for methyl salicylate (No. 899). The Committee noted that the estimated daily intake of 0.7 mg/kg bw of methyl salicylate, based on poundage data, is approximately equal to its ADI of 0–0.5 mg/kg bw within the precision of the intake estimates. The Committee received an analysis of intake based on individual dietary records for methyl salicylate for its potential use in mint-flavoured baked goods, chewing-gum, hard and soft sweets, and beverages. This analysis showed that > 50% of the population would be expected to consume methyl salicylate. The (appropriate) use of this measured proportion of eaters in place of the default assumption of 10% yields an estimated intake of methyl salicylate of 0.1 mg/kg bw, which is less than the current ADI of 0–0.5 mg/kg bw.

On the basis of the available data on metabolism and toxicity, the Committee concluded that none the flavouring agents in this group would present a safety concern when used at their current levels. Other data on toxicity, including the results of studies on genotoxicity and developmental toxicity, were consistent with the results of the safety evaluations conducted according to the Procedure.

The background information relevant to a safety evaluation of 46 hydroxy- and alkoxy-substituted benzyl derivatives used as flavouring agents are summarized below and in Tables 1 and 2.

Production volumes and intake values for each flavouring agent in this group are shown in Table 2.

Twenty-nine of the 46 flavouring agents have been detected as natural components of traditional foods (Maarse et al., 1999; Table 2). Quantitative data on natural occurrence and consumption ratios have been reported for 13 substances in the group. The consumption of eight substances is derived predominantly from their presence in traditional foods (i.e. they have a consumption ratio > 1), whereas para-methoxybenzaldehyde (No. 878), methyl-ortho-methoxybenzoate (No. 880), vanillin (No. 889), methyl salicylate (No. 899), and ethyl salicylate (No. 900) are not consumed primarily from traditional foods (consumption ratio, < 1).

2.3.1 Biochemical data

The hydroxy- and alkoxy-substituted benzyl derivatives have been shown to be rapidly absorbed in the gastrointestinal tract, metabolized in the liver to yield benzoic acid derivatives, and excreted primarily in the urine either unchanged or conjugated (Jones et al., 1956; Davison, 1971). Metabolites participate in enterohepatic cycling to some extent, leading to further metabolism by gut bacteria.

In rabbits, 96% of a single oral dose of 400 mg/kg bw 4-hydroxybenzaldehyde (No. 956) was excreted in the urine within 24 h as 4-hydroxybenzoic acid and its glycine, glucuronic acid, and sulfate conjugates (Bray et al., 1952).

In a similar study, groups of four to eight rabbits were given 4-hydroxybenzoic acid (No. 957) at a dose of 100, 250, 500, 1000, or 1500 mg/kg bw by gavage every 3–7 days. Urine was collected continuously and analysed for metabolites. The total urinary recovery of the test material ranged from 84% to 104% . Glucuronic acid and sulfate conjugates were also detected in the urine, at 10–35% and 4–7%, respectively. The concentrations of all the metabolites returned to background values within 24 h after dosing (Bray et al., 1947). In a corresponding study, approximately 94% of 2-hydroxybenzoic acid (No. 958) at a single oral dose of 250 or 500 mg/kg bw given to two groups of four rabbits was excreted unchanged or as the glucuronic acid and sulfate conjugates (Bray et al., 1948).

About 6% of a dose of 52 mg of 2,4-dihydroxybenzaldehyde (No. 908) given by intraperitoneal injection to female albino rats was excreted in the urine as the corresponding hippurate within 24 h (Teuchy et al., 1971).

Three patients being treated for rheumatic fever were given an oral dose of 5330–6000 mg of 2,4-dihydroxybenzoic acid as 1000 mg per dose every 3 h for 2–16 days. The average daily rate of urinary excretion was 43–76%. The average daily excretion of sulfate conjugate per patient was essentially constant throughout the study, but the average daily excretion of glucuronic acid conjugate increased by four- to sixfold over the 16 days (Clarke et al., 1958). In an investigation of the presence of dihydroxybenzoic acid isomers in the urine of 15 persons, only 3,5-dihydroxybenzoic acid was detected (Williams, 1965).

Groups of three or more fasted dogs were given butyl para-hydroxybenzoate (No. 870) at a dose of 1000 mg/kg bw orally or 50 mg/kg bw intravenously. Blood and urine samples were collected at fixed intervals until the concentrations returned to background values within 48 h. The test material was recovered almost entirely as the para-hydroxybenzoic acid conjugate of glucuronic acid, at 48% after the oral dose and 40% after the intravenous dose. Most of the material was excreted between 6 and 30 h after dosing. Although the relatively low rate of recovery with both methods of administration was attributed to incomplete hydrolysis of the ester in the body, incubation of the butyl ester with freshly prepared liver homogenate in vitro showed complete hydrolysis within 30–60 min. In studies with related benzoate esters, such as methyl and ethyl para-hydroxybenzoate, significantly larger amounts of material were recovered (Jones et al., 1956). This finding suggests that increased amounts of the homologous series of alkyl esters may activate other metabolic and excretion pathways. The authors concluded that butyl para-hydroxybenzoate and other alkyl esters are readily absorbed, metabolized, and excreted.

Ten rabbits were each fed 200 mg of veratraldehyde (No. 877) by stomach tube, and urine was collected over the next 24 h. About 70% of the material was recovered in the urine as free corresponding acid (~28%) and its glucuronic acid (~38%) or sulfate (3–7%) conjugate (Sammons & Williams, 1941).

Because of its prevalence and importance as a flavouring agent, vanillin (No. 889) has been the subject of numerous studies on metabolism. Male albino rats were given 100 mg/kg bw of vanillin in a solution of propylene glycol and water by stomach tube; urine and faeces were collected separately for 24-h periods, and bile samples were collected by cannulation of the common bile duct. Only trace amounts of benzoic acid derivatives remained in the urine after the first 24 h, and none remained after 48 h. Free and conjugated forms of vanillic acid and vanillyl alcohol accounted for 94% of the dose in the urine. Vanillin and its primary reduction and oxidation metabolites were also excreted in appreciable amounts in the bile. Bile collected for 5 h from two rats given vanillin at a dose of 100 or 300 mg/kg bw orally contained glucuronide conjugates of vanillin (6%), vanillyl alcohol (8%), and vanillic acid (9%) (Strand & Scheline, 1975).

In Sprague-Dawley albino rats, 60% of a dose of vanillin of 100 mg/kg bw in 0.9% NaCl given by intraperitoneal injection was recovered in the 24-h urine. The urinary metabolites included unconjugated vanillic acid, the sulfate and glucuronic acid conjugates of vanillic acid, and conjugates of vanillyl alcohol, vanillin, and catechol. The urinary glycine conjugate of vanillic acid was not found in this study (Wong & Sourkes, 1966).

Three rabbits given vanillin by gavage at 1000 mg/kg bw excreted an average of 83% of the dose in their urine, with 69% of the dose as free and conjugated vanillic acid and 14% as conjugated vanillin (Sammons & Williams, 1941).

An adult person received 100 mg of vanillin dissolved in water, and urine was collected for 24 h. The concentration of vanillic acid in the urine increased from a background value of 0.3 mg per 24 h to 96 mg per 24 h, accounting for about 94% of the dose (Dirscherl & Wirtzfeld, 1964).

After administration of a single dose of 400 mg/kg bw of salicylaldehyde (No. 897) to a fasted rabbit, 75% of the dose was excreted in urine collected over 24 h. Urine analysis revealed mainly ether-soluble acids, 27% and 3% being accounted for by glucuronic acid and sulfate conjugates of vanillic acid, respectively (Bray et al., 1952).

In male rats, 94% of a dose of 150 mg/kg bw of piperonal (No. 896) in propylene glycol administered by gavage was accounted for in the urine within 24 h. No unchanged compound was excreted, and no metabolites were detected in the urine more than 48 h after dosing (Klungsoyr & Scheline, 1984). In a study of the metabolism of the corresponding esters piperonyl acetate (No. 894) and piperonyl isobutyrate (No. 895) given at 100 mg/kg bw by gavage to male rabbits, 70% of the dose of piperonyl acetate and 11% of that of piperonyl isobutyrate was recovered in 72-h urine (Wright & Holder, 1980).

The results of these studies indicate that the hydroxy- and alkoxy-substituted benzyl derivatives are rapidly absorbed, metabolized, and excreted in the urine mainly as sulfate and glucuronic acid conjugates of the corresponding hydroxybenzoic acid derivatives.

Aromatic esters can be expected to be hydrolysed in vivo by the catalytic activity of carboxylesterases or esterases (Heymann, 1980). B-Esterases, the most important of the group, are active in most mammalian tissue (Heymann, 1980; Anders, 1989) but predominate in hepatocytes (Heymann, 1980). Acetals hydrolyse uncatalysed in gastric juice and intestinal fluids to yield the corresponding aldehydes. In vivo, substituted benzyl esters and benzaldehyde acetals are hydrolysed to the corresponding alcohols, aldehydes, and carboxylic acids.

In 3- to 6-month-old male rabbits, 83% of an oral dose of 100 mg/kg bw of piperonyl acetate (No. 894) and 15% of a aimilar dose of piperonyl isobutyrate (No. 895) were hydrolysed and excreted as either free or conjugated piperonylic acid within 72 h; < 1% of piperonyl alcohol was excreted (Wright & Holder, 1980).

An oral dose of methyl salicylate (No. 899) equivalent to 500 mg/kg bw of salicylic acid was dissolved in 2% methyl cellulose and given to male rats. The plasma concentrations measured within 20 min of dosing showed complete hydrolysis of methyl salicylate. In a similar experiment with three fasted male dogs given 320 mg/kg bw of methyl salicylate in capsules, blood drawn 1 h after dosing showed 95% hydrolysis of methyl salicylate to salicylic acid. In six persons, 79% of a dose of 0.42 ml (approximately 500 mg) of methyl salicylate administered in ginger ale was hydrolysed in the blood within 90 min (Davison et al., 1961).

Several experiments were conducted to study the metabolism of esters of para-hydroxybenzoic acid after oral (1000 mg/kg bw) or intravenous (50 mg/kg bw) administration to dogs. The esters were absorbed in the gastrointestinal tract and rapidly hydrolysed by esterases in the liver and kidney. In the case of butyl para-hydroxybenzoate (No. 870), 48% was recovered after oral and 40% after intravenous administration. Liver preparations from dogs injected with 100 mg/kg bw of the methyl, ethyl, or propyl ester showed 100% hydrolysis within 3 min. In the case of the butyl ester, 100% hydrolysis occurred within 30–60 min (Jones et al., 1956).

Benzyl acetate was rapidly hydrolysed to benzyl alcohol in vitro, with a peak alcohol concentration after 4 min. The absence of benzyl acetate in plasma indicates that benzyl acetate is rapidly hydrolysed to benzyl alcohol, which is then rapidly oxidized to benzoic acid (Yuan et al., 1995). In vitro, 90% of benzyl phenylacetate was hydrolysed within 1 h and 100% within 2 h of incubation with a 2% pancreatin solution (Leegwater & van Straten, 1974).

Acetals of benzaldehyde are also readily hydrolysed. Benzaldehyde propylene glycol acetal was 97% hydrolysed after incubation for 5 h with simulated gastric juice and intestinal fluid in vitro (Morgareidge, 1962).

In general, hydroxy- and alkoxy- derivatives of benzaldehyde and benzyl alcohol are oxidized to the corresponding benzoic acid derivatives and, to a lesser extent, reduced to the corresponding benzyl alcohol derivatives. The resulting hydroxy- and alkoxy-benzoic acid derivatives form sulfate, glucuronic acid, or glycine conjugates, depending mainly on ring substitution. Hydroxy- and methoxy-substituted benzoic acid derivatives such as vanillic acid tend to form sulfate or glucuronic acid conjugates, while methylenedioxy-substituted benzoic acid derivatives such as piperonylic acid form glycine conjugates. Benzoic acid hydroxy- and alkoxy-derivatives undergo decarboxylation and O-demethylation to a minor extent. Protocatechuic acid is a key intermediate formed by O-demethylation of benzoic acid (Wong & Sourkes, 1966; Strand & Scheline, 1975). Benzyl alcohol derivatives may also be reduced in gut microflora to toluene derivatives, especially if a free para-hydroxyl group is present (Strand & Scheline, 1975; see Figure 1).

In a study of the degradation of some methoxylated aromatic compounds by Actinomyces aureus A-94, anisyl alcohol (No. 871) was oxidized to the corresponding anisic acid, then demethylated and hydroxylated to yield protocatechuic acid (3,4-dihydroxybenzoic acid). Additional enzymatic action converted protocatechuic acid to succinic acid via beta-carboxymuconic acid and beta-oxoadipic acid. In the final stage, succinic acid entered the tricarboxylic acid cycle (Tsai et al., 1965).

Analysis of the medium after incubation of anisyl alcohol (No. 871) with rat caecal extract after approximately 46 h showed the presence of anisic acid. O-Demethylation was not observed (Scheline, 1972).

In the investigation of the metabolism of various aromatic aldehydes and alcohols by rat intestinal microflora, the major metabolites were products of reduction. Microflora-mediated metabolic transformations included reduction, dehydroxylation, O-demethylation, and decarboxylation, leading to a variety of derivatives of benzyl alcohol, benzoic acid, and toluene (Scheline, 1972). After incubation of vanillyl alcohol (No. 886) with rat caecal extract, vanillic acid and the toluene derivatives, 4-methylguaiacol and 4-methylcatechol from complete reduction of the alcohol functional group were present (Scheline, 1972). Analysis of urinary metabolites collected over the first 24 h from male rats given vanillyl alcohol by gavage at a dose of 100 or 300 mg/kg bw showed the presence of vanillic acid and traces of vanillyl alcohol and the glycine conjugate of vanillic acid. Smaller quantities of conjugated fractions of vanillin, guaiacol, catechol, 4-methylguaiacol, and 4-methylcatechol were also found. The presence of catechol and 4-methylcatechol indicated that decarboxy-lation and complete reduction of the alcohol function, respectively, occurred in vivo (Strand & Scheline, 1975).

Rabbits were given 2000 mg/kg bw of veratraldehyde (No. 877) by oral gavage, and urine was collected for 24 h. About 70% of the aldehyde was accounted for in urine, mainly as the corresponding acid veratric acid (28%), and its glucuronic acid conjugate (38%). Veratric acid was decarboxylated to a small extent and O-demethylated to yield catechol (Sammons & Williams, 1946). Presumably, veratric acid enters the enterohepatic circulation where gut microflora decarboxylate it to yield catechol (ortho-hydroxyphenol). Formation of catechol after incubation of veratraldehyde with rat caecal extract indicates that this decarboxylation pathway exists in gut bacteria (Scheline, 1972).

Analysis of the metabolites produced 46 h after incubation of anisaldehyde (No. 878) with rat caecal preparations revealed the presence of anisic acid and anisyl alcohol. Anisaldehyde thus undergoes oxidation and reduction in caecal preparations (Scheline, 1972). In rabbits, about 75% of an oral dose of 2000 mg/kg bw of anisaldehyde was excreted as the glucuronic acid conjugate of para-methoxybenzoic acid (anisic acid) within 24 h (Sammons & Williams, 1946).

In groups of four to six male rats, a single oral dose of 100 or 300 mg/kg bw of vanillin (No. 889) was metabolized and excreted in urine as vanillin, vanillic acid, and vanillyl alcohol within the first 24 h. Small amounts of O-demethylated, decarboxylated, and further reduced metabolites were also identified, which included protocatechuic acid (3,4-dihydroxybenzoic acid, product of O-demethylation), guaiacol (ortho-methoxyphenol, product of decarboxylation), vanilloylglycine, catechol (ortho-hydroxyphenol), 4-methylguaiacol (product of alcohol functional group reduction), and 4-methylcatechol (product of reduction and ring hydroxylation). Only traces of vanillic acid derivatives were detected in urine collected between 24 and 48 h, and no metabolites were detected in urine collected between 48 and 96 h. In a similar experiment in rats, more than 94% of a single oral dose of 100 mg/kg bw of vanillin was excreted in urine within 48 h. Between 65 and 70% of the urinary metabolites were oxidation products (Strand & Scheline, 1975).

Most of a dose of 100 mg/kg bw of vanillin administered by intraperitoneal injection to rabbits was excreted in urine 24 h later. About 69% was oxidized to vanillic acid, and 10% was reduced to vanillyl alcohol. More than 10% was excreted as the glucuronic acid conjugate of vanillin (Sammons & Williams, 1941).

In rats, less than 6% of a dose of 52 mg of 2,4-dihydroxybenzoic acid given by intraperitoneal injection was excreted as the corresponding hippurate (Teuchy et al., 1971). In persons given daily oral doses of up to 6000 mg of 2,4-dihydroxybenzoic acid for up to 16 days, the main urinary metabolites were the glucuronic acid and sulfate conjugates (Clarke et al., 1958).

In albino male Wistar rats, 94% of an oral dose of 150 mg/kg bw of piperonal (No. 896) and 90% of a dose of 150 mg/kg bw of piperonyl alcohol dissolved in propylene glycol were excreted in the urine within 24 h. The main urinary metabolites of both substances were piperonylic acid (17–20%) and the glycine conjugate of piperonylic acid (71–72%). No unchanged compound was excreted, and no metabolites were detected 48 h after dosing. Less than 0.7% of the dose was accounted for by O-demethylenation to protocatechuyl alcohol, protocatechu-aldehyde and protocatechuic acid (Klungsoyr & Scheline, 1984).

In male Swiss-Webster mice, 87–93% of an oral dose of 0.75 mg/kg bw of [methylene-¹⁴C]piperonyl alcohol, piperonal, or piperonylic acid administered in dimethyl sulfoxide (50 µl followed by a wash with 100 µl) was accounted for in the urine within 48 h. Most was eliminated within the first 12 h. Less than 10% was excreted in the faeces. In all cases, the major metabolite was the glycine conjugate of piperonylic acid. Minor amounts of unchanged piperonylic acid were also present (Kamienski & Casida, 1970).

Esters of hydroxy- and alkoxy-substituted benzyl derivatives can be expected to be hydrolysed to the corresponding benzyl alcohol and benzoic acid derivatives, while the acetals are hydrolysed to the parent benzaldehyde derivative. Thus formed, the alcohols and aldehydes are oxidized mainly to benzoic acid derivatives, which are either excreted unchanged or form sulfate, glycine, or glucuronic acid conjugates. Some glucuronic acid conjugates may pass into the bile and enter the enterohepatic circulation, where they are hydrolysed or subjected to the reduction reactions of gut bacteria. Small amounts of hydroxy- and alkoxybenzoic acid derivatives have been reported to undergo reductive decarboxylation in the gut. Other minor metabolic detoxication pathways include O-demethylation and ring hydroxylation.

2.3.2 Toxicological studies

The LD₅₀ values after oral administration for 38 of the 46 benzyl derivatives ranged from 520 mg/kg bw to 13 000 mg/kg bw in male and female rats, guinea-pigs, mice, and rabbits (Deichmann & Kitzmiller, 1940; Draize et al., 1948; Sokol, 1952; Giroux et al., 1954; Doull et al., 1962; Jenner et al., 1964; Taylor et al., 1964; Hagan et al., 1965; Fogleman & Margolin, 1970; Fujii et al., 1970; Davison et al., 1961; Weir & Wong, 1971; Moreno, 1973; Sado, 1973; Moreno, 1974; Wohl, 1974; Levenstein, 1975; Grady et al., 1976; Moreno, 1976, 1977; BASF, 1981; Givaudan Corp., 1982; Mondino, 1982; Moreno, 1982; Peano & Berruto, 1982; Mallory et al., 1983; Sterner & Chibanguza, 1983; National Toxicology Program, 1984; Ohsumi et al., 1984; Ohta et al., 1984; Reagan & Becci, 1984; Inouye et al., 1988; Buch, 1989; Hasegawa et al., 1989; Cerven, 1990; Dow Chemical Co., 1992; Dufour, 1994; Sanders & Crowther, 1997). The LD₅₀ values of most of the compounds are > 1000 mg/kg bw (see Table 3).

Short-term studies of toxicity have been performed with eight of the substances in this group, but most with butyl para-hydoxybenzoate (No. 870), commonly known as butyl paraben, vanillin (No. 889), methyl salicylate (No. 899), ethyl vanillin (No. 893), and piperonal (No. 896). Each of these four substances has been allocated an ADI. The studies with ethyl vanillin were reviewed comprehensively by the Committee at its thirty-fifth and forty-sixth meetings (Annex 1, references 89 and 117), and the results of that review are included below so as to present all relevant data related to the safety of structurally related hydroxy- and alkoxy-benzaldehyde derivatives.

Groups of 10 male and 10 female 8-week-old ICR/jcl mice were maintained on diets containing pelletized butyl-para-hydroxybenzoate at concentrations calculated to provide an average intake of 900, 1900, 3800, 7500, or 15 000 mg/kg bw per day for 6 weeks. Twenty male and 20 female mice of the same age and strain were used as controls and fed the basal diet. Survival, body-weight gain, and histolological end-points were measured. All the animals given 7500 or 15 000 mg/kg bw per day died within the first 2 weeks of treatment. The body-weight gain of those given 1900 or 3800 mg/kg bw per day was less than 10% of that of the control group, whereas the animals at the lowest concentration showed weight gain similar to that of controls. Histological examination showed atrophy of lymphoid tissue and liver degeneration and necrosis in all groups except those at the lowest dose (Inai et al., 1985).

Butyl-para-hydroxybenzoate was dissolved in soya bean oil at a concentration of 100 mg/0.5 ml and administered by oral intubation to rats for 13–15 weeks at concentrations calculated to result in a daily intake of 0, 0.25, or 50 mg/kg bw. Body weight, measured twice weekly, showed no significant difference from controls. Some animals were killed on a predetermined schedule for histological evaluation. There were no sporadic deaths and no significant histological differences from controls. The NOEL was 50 mg/kg bw per day (Ikeda & Yokoi, 1950).

Groups of 12 Wistar rats of each sex were fed a powdered mixture of butyl-para-hydroxybenzoate and dog chow providing a dose of 0, 2000, or 8000 mg/kg bw per day for 12 weeks. Body weight and food intake were measured every 2 weeks, and necropsy and histological examinations were performed at the end of the experiment. Animals found dead before the end of the study were necropsied and the appropriate tissues were fixed for histopathological evaluation. Although there were no effects at the low dose, all the males and many of the females at the high dose died within several weeks of the beginning of treatment. The body weights and motor activity of these animals were decreased, and they had a slower growth rate than controls. The NOEL was 2000 mg/kg bw per day (Matthews et al., 1956).

A group of 10 male and 10 female weanling rats received a diet containing a mixture of eugenol (90 mg/kg bw), para-methoxybenzaldehyde (7.3 mg/kg bw), and piperonal (16 mg/kg bw) at a concentration calculated to provide an average daily intake of 110 mg/kg bw for 90 days. Food intake and growth were measured each week. After 12 weeks on the test material, sugar, albumin, and haemoglobin concentrations were measured in the urine of three male and three female animals. No sugar was found; although the urine of male rats contained albumin, the authors considered the finding pathologically insignificant. All surviving animals were killed at 90 days and necropsied, and organs were weighed. Weekly measurements of body-weight gain, food intake, haematological and clinical chemical end-points, and gross and histological examinations showed no significant difference between test and control animals. The NOEL was 7.3 mg/kg bw per day for para-methoxy-benzaldehyde (No. 878) and 16 mg/kg bw per day for piperonal (No. 896) (Trubeck Laboratories Inc., 1958).

Groups of five male and five female 8-week-old rats were fed diets containing para-methoxybenzaldehyde (No. 878) and piperonal (No. 896) at a concentration providing 500 mg/kg bw per day for 16 weeks, and groups of five male and five female 4-week-old rats were fed the same substances at 50 mg/kg bw per day for 28 weeks. One control group of 20 rats was used for each experiment. No gross pathological changes were seen in any group. The NOEL was 500 mg/kg bw per day for both para-methoxybenzaldehyde and piperonal (Food & Drug Administration, 1954).

Groups of five male and five female weanling Osborne-Mendel rats were given diets containing para-methoxybenzaldehyde (No. 878) or piperonal (No. 896) at a concentration providing 50 mg/kg bw per day for 27–28 weeks or 500 mg/kg bw per day for 15 weeks. Ten male and 10 female rats were used as controls. Body weight, food intake, and general condition were noted weekly, and haematological examinations were performed at the end of the study. No effects were seen at either dose. The NOEL was 500 mg/kg bw per day for both para-methoxybenzaldehyde and piperonal (Hagan et al., 1967).

Weanling Osborne-Mendel rats (number per group unspecified) were fed a diet containing methyl salicylate at concentrations calculated to provide an average daily intake of 500 mg/kg bw for 16 weeks and 50 mg/kg bw for 28 weeks. No adverse effects were observed at either dose (Hagan et al., 1965).

Groups of five male and five female Fischer 344 rats, 28 days old, were maintained on diets containing methyl ortho-methoxybenzoate in corn oil at a concentration calculated to provide an average daily intake of 0 or 94 mg/kg bw for 14 days. The animals were observed for deaths twice a day, and body weight and food consumption were measured weekly. No significant differences were found between test and control groups. At necropsy, the liver and kidney were weighed and examined histologically. The only significant result was a decrease in relative kidney weight in males, but the absolute liver and kidney weights were unchanged in treated animals, and no kidney lesions were found; the authors concluded that the change was not related to treatment. The NOEL was 94 mg/kg bw per day (Van Miller & Weaver, 1987).

The protocols used by the Food & Drug Administration and by Hagan et al. to study para-methoxybenzaldehyde and piperonal were also used in studies in which Osborne-Mendel rats were maintained on a diet containing vanillin at a concentration calculated to provide an average daily intake of 500 mg/kg bw for 16 weeks. In the study by Hagan et al., an additional group of 10 rats was given 50 mg/kg bw per day of vanillin for 28 weeks. No effects were seen at either dose in either study (Food & Drug Administration, 1954; Hagan et al., 1967).

Vanillin or ethyl vanillin was dissolved in corn oil and added to the diet of five male weanling Osborne-Mendel rats at concentrations calculated to provide an average daily intake of 1000 or 2500 mg/kg bw for 1 year. Ten male and 10 female rats were fed a diet containing 3% corn oil as a control. Weekly measurements of body weight and food intake and observations of general condition showed no differences between test and control groups. No differences in haematological parameters were seen at necropsy. The NOEL for vanillin and ethyl vanillin was 2500 mg/kg bw per day (Hagan et al., 1967).

Four groups of eight young albino rats were fed vanillin or ethyl vanillin as a 4% solution in milk at an estimated daily intake of either 20 mg/kg bw for 126 days or 64 mg/kg bw for 70 days. In the 70-day study, half the animals were killed and the other half were put on a recovery diet for 8 more weeks. Additionally, 12 rats were given a dose of 300 mg/kg bw of vanillin or ethyl vanillin as a 4% solution in olive oil orally by gavage twice per week for 14 weeks. Observation of appearance, behaviour, and body-weight gain showed a reduced growth rate and myocardial, renal, hepatic, lung, spleen, and stomach injuries at the dose of 64 mg/kg bw (nature not specified) (Deichmann & Kitzmiller, 1940).

Groups of 20 male and 20 female CD Sprague-Dawley rats were given ethyl vanillin at a dose of 0, 500, 1000, or 2000 mg/kg bw per day in the diet for 13 weeks. Food consumption and body weight were recorded weekly; ophthalmoscopy was performed before treatment and at the end of the study; detailed haematological and clinical chemical examinations were carried out at weeks 6 and 13. At termination, all the animals were necropsied and their organs were weighed. A complete histological examination was performed on controls and rats at the highest dose and was extended to animals at the the low and intermediate doses when treatment-related effects were suspected.

Some differences in food intake and body-weight gain were found in comparison with controls, but the authors considered these to be unrelated to treatment because their severity was not dose-related and intragroup variation in feeding patterns was seen among male rats. No treatment-related differences in haematological parameters were seen at week 6 or at termination.

At autopsy, enlarged cervical lymph nodes were seen in males at the intermediate dose and in animals of each sex at the highest dose. In addition, rats of each sex at the highest dose showed a reduction in adipose tissue. The absolute liver weights were similar to those of controls, but the relative liver weights were increased in rats at the intermediate and high doses. An increased incidence of hepatic peribiliary inflammatory change was seen in males and females at the two higher doses and minor bile-duct hyperplasia in a few males at these doses. No changes were observed in the liver parenchyma and no degenerative or inflammatory changes of the bile-duct epithelium were found. The NOEL was 500 mg/kg bw per day (Hooks et al., 1992).

A series of studies was performed by Abbott & Harrisson (1978) to evaluate the effect of methyl salicylate on bone density. Groups of five male and five female rats were maintained on diets containing methyl salicylate at concentrations calculated to provide an average daily intake of 0, 100, 180, 320, 560, or 1000 mg/kg bw for 11 weeks. The groups received the substance at 50% of the final dose for the first 2 weeks, 75% for the third and fourth weeks, and 100% thereafter. Body weight and food consumption were measured weekly. Males at 320 mg/kg bw per day showed decreased weight gain when compared with controls, and males and females at 560 and 1000 mg/kg bw per day showed both decreased weight gain and increased bone density at the metaphyses of the femur, humerus, tibia, and radius. No such changes were observed at lower doses.

In a 12-week study, two groups of five male rats were fed diets containing methyl salicylate at concentrations calculated to provide an average daily intake of 300 or 1000 mg/kg bw. No control animals were used. All the animals at the high dose died during the first 6 weeks of the study, and were found to have bone lesions on full-body X-rays. All animals at the low dose survived to the end of the study, and none had the bone lesions observed at the high dose.

In a second 11-week study to re-evaluate the effects of a group of structurally related substances on bone density, methyl salicylate was administered in the diet to provide a dose of 1000 mg/kg bw per day to 15 rats. Twenty rats served as controls. Whole body X-rays were performed periodically, and body weights were checked. The mortality rate in the experimental group was 20%, whereas all the control animals survived. Changes in bone density similar to those seen in the first study were reported with both methyl salicylate and acetylsalicylic acid, indicating that the bone changes were associated with high doses of ortho-hydroxybenzoic acid derivatives (salicylates).

In a 6-week study, animals were fed a diet ad libitum containing methyl salicylate estimated to provide an average daily intake of 0, 300, or 1000 mg/kg bw. Ten paired feeding groups were given diets providing methyl salicylate at 300 mg/kg bw per day, methyl para-hydroxybenzoate at 1000 mg/kg bw per day, or no treatment. All pair-fed animals received daily rations ad libitum, equal to the mean daily amount of food consumed by the group given 1000 mg/kg bw per day methyl salicylate. The group given 300 mg/kg bw per day had a slightly lower growth rate than the controls, whereas the changes in body weight of the pair-fed groups were similar to those of rats given methyl salicylate at 1000 mg/kg bw per day ad libitum. Deaths occurred among rats at the high dose of methyl salicylate ad libitum and in all members (including the control group) of the pair-fed groups.

A study was undertaken to evaluate the progression of bone change, in which groups of 10 male and 10 female rats were fed diets providing a dose of methyl salicylate of 300, 450, 600, or 1000 mg/kg bw per day for 11 weeks. A fifth group was fed the basic diet and used as controls. Two animals per group were X-rayed weekly and killed a week later until the end of the experiment. The femurs of certain animals were examined histologically. The group at the highest dose began showing bone changes at week 2, and those at 600 mg/kg bw per day group began showing signs at week 5. The X-rays of the remaining groups were normal throughout the 11-week study. Histological analysis showed an increased incidence of cancellous bone in rats at the high dose at week 2 and in those at 600 mg/kg bw per day by week 8. The NOEL was 450 mg/kg bw per day.

Two groups of 10 male and 10 female rats were maintained on a diet designed to result in an average daily intake of 0, 50 or 500 mg/kg bw of methyl salicylate for 17 weeks. Weekly measurements of body weight and of organ weights at necropsy and the results of gross and microscopic examinations showed no adverse effects (Webb & Hansen, 1963).

Three male and three female rats were fed a diet containing methyl salicylate at a concentration resulting in a dose of 0 or 1000 mg/kg bw per day. As each experimental animal died, one control animal was killed at the same time until all of the animals were dead. The first male died at 11 days and two more at 19 days. The first female died at 31 days, the second at 40 days, and the third at 71 days. The last animal to die was examined by X-rays and microscopically. The treatment was not only lethal but increased the bone density in the metaphysis of all bones and induced laboured respiration, and gastric haemorrhages in the glandular stomach (Webb & Hansen, 1963). The effect of salicylates on bone density was also reported in an abstract (Harrisson et al., 1963).

Isoamyl salicylate was incorporated in the diet of groups of 15 male and 15 female Wistar rats at a concentration of 0, 50, 500, or 5000 mg/kg for 13 weeks. On the basis of data on body weight and food intake, the authors calculated mean intakes of 0, 4.7, 46, and 420 mg/kg bw per day for males and 0, 4.8, 47, and 480 mg/kg bw per day for females throughout treatment. An additional group of five females and five males received only the two higher doses for 2 or 6 weeks. The animals were weighed on days 1, 2, 6, 9, and 13 and then weekly until day 91. Food and water consumption was measured for 24 h before weighing, and urine was collected during the last 2 days of treatment and analysed. At necropsy, organs were weighed and haematological and histopathological examinations were performed. Rat at the highest dose showed increased relative kidney weights, and males at 25 mg/kg bw per day had increased relative and absolute kidney weights, although with no associated histopathological changes. Decreased weight gain and food intake were observed at the highest dose. In a separate paired-feeding study in which two groups of 10 rats were fed either 0 or the highest dose of methyl salicylate for 98 days, decreases in body weight and food consumption were considered by the authors to be due to poor palatability of the diet. The NOEL was 4.7 mg/kg bw per day (Drake et al., 1975).

Vanillin and ethyl vanillin were given as a solution in milk to one rabbit at 240 mg/kg bw per day for 56 days and to two rabbits at the same dose for 126 days. Both substances were also administered as a solution in 10% glycerol, vanillin at a dose of 83 mg/kg bw per day for 14 days or 103 mg/kg bw per day for 61 days and ethyl vanillin at a dose of 15 mg/kg bw per day for 15 days, 15 mg/kg bw per day for 31 days, 32 mg/kg bw per day for 17 days, 41 mg/kg bw per day for 31 days, or 49 mg/kg bw per day for 49 days.

Appearance, behaviour, and body-weight gain were not significantly affected. There were no gross or histopathological alterations in test animals. In rabbits receiving the substances in glycerol solution, anaemia, diarrhoea, and lack of weight gain were observed at the highest dose; no toxic effects were seen at lower doses. Glycerol poisoning, evidenced by restlessness, tremor, convulsions, and coma, was observed in rabbits given 83 mg/kg bw per day of vanillin for 14 days and in those given 15 mg/kg bw per day of ethyl vanillin for 15 days (Deichmann & Kitzmiller, 1940).

Four groups of three male and three female purebred beagles were given capsules containing methyl salicylate at a dose of 150, 300, 500, or 800 mg/kg bw per day for 7.5 months. The dogs were given one-half of the dose at the morning feed and the other half in the afternoon for 6 days/week. Two males and four females were fed the same diet without methyl salicylate and served as a controls. After 6.5 months, three dogs at 300 mg/kg bw per day, two at 150 mg/kg bw per day, and all the control animals were killed, and three at 300 mg/kg bw per day were put on a recovery diet. After 7 months, all the remaining animals were killed, except those returned to the normal diet, which were killed after 8 months.

Body weight was measured weekly and haematological examinations were performed at the fifth month on animals at the two lower doses. After sacrifice, the organs were weighed, and gross and histological examinations were performed. All the animals at the highest dose were dead by the second week. Of the groups at 500 mg/kg bw per day, only two survived the duration of the experiment, while the rest died at weeks 2, 3, 5, and 8. The body weights of the groups at the two lower doses were not significantly different from those of controls, and haematological examinations showed normal values. Increased liver and kidney weights were seen at all doses but not in the animals placed on the recovery diet.

In a further experiment with the same protocol, dogs were fed capsules containing methyl salicylate at a dose of 50 (four males, four females), 100, or 170 mg/kg bw per day (six males, six females). All animals were killed at 6 months, except for two animals of each sex at the highest dose and four controls, which were placed on a recovery diet for 2 months. The same measurements were made as in the preceding study, except that only liver and kidney weights were assessed. There were no adverse effects at any dose. The NOEL was 170 mg/kg bw per day (Abbott & Harrisson, 1978).

Six groups of one male and one female dog were given a capsule containing methyl salicylate at a dose of 0, 50, 100, 250, 500, 800, or 1200 mg/kg bw per day on 6 days/week for up to 59 days. The three animals at the highest dose were either killed in extremis or died within the first month of the experiment. At necropsy, one dog at 800 mg/kg bw per day and both at 250 mg/kg bw per day were examined grossly and microscopically. The dogs at the higher dose vomited 3–4 h after each dose throughout the study. Those at 500 mg/kg bw per day had diarrhoea and weakness during the last 3–4 days of the study. Both animals at the highest dose and one at 800 mg/kg bw per day showed marked fatty metamorphosis in the liver (Webb & Hansen, 1963).

Seven long-term studies of toxicity were performed with five of the flavouring agents in this group. Two 2-year studies were conducted on methyl salicylate (No. 899) in rats and one in dogs (Packman et al., 1961; Webb & Hansen, 1963).

Three groups of 50 male and 50 female ICR/jcl mice, 8 weeks old, were fed pellets containing butyl-para-hydroxybenzoate at a dose of 0, 225, 450, or 900 mg/kg bw per day for 102 weeks. Intake was measured once a week for the first 30 weeks, once every other week for the following 20 weeks, and then once every 4 weeks until the end of the experiment. Body weight was measured once a week for the first 6 weeks, once every other week for the next 24 weeks, and then every 4 weeks for the duration of the experiment. All animals found dead during the experiment were necropsied, and all those that were still alive at the end of the experiment were killed and necropsied at week 106. The tissues of all animals, regardless of time of death, were examined histologically. The tumour incidence in the experimental groups was not significantly different from that in the control group. The NOEL was 900 mg/kg bw per day (Inai et al., 1985).

Vanillin or ethyl vanillin dissolved in propylene glycol was added to the diet of groups of 12 male and 12 female rats at a concentration estimated to provide an average daily intake of 250, 500, or 1000 mg/kg bw, for 2 years. Twenty control rats were fed 3% propylene glycol. Weekly measurements of body weight and food intake and observations of general condition failed to show any differences between test and control groups. Haematological examinations at necropsy showed no effects in any of the animals at any concentration. The NOEL for vanillin and ethyl vanillin was 1000 mg/kg bw per day (Hagan et al., 1967).

Piperonal was incorporated in the diet of male and female rats at a concentration of 0.5 or 1.1% for 2 years. The appearance, behaviour, and weight of the animals under study were noted, and histological examinations were conducted. The authors reported that no specific results were obtained (Bär & Griepentrog, 1967). As the report of this study that was available was incomplete and difficult to read, these data were not considered in detail.

Groups of 25 male and 25 female rats were fed a diet calculated to provide a dose of methyl salicylate of 0, 50, 250, 500, or 1000 (24 males, 26 females) mg/kg bw per day for 2 years. The diet was prepared every 2 weeks, and an additional 10% methyl salicylate was added to each mix in order to account for evaporation. The animals were weighed weekly. Haematological examinations performed at 3, 11, 17, and 22 months on 10 rats from each group showed normal values. Animals in the control group and at the two higher doses were examined histologically. Gross lesions, organ weights, bones, and muscles were assessed upon termination in each group. Half of the rats at the highest dose died after week 8. Only five rats survived past week 20 and one past week 35, finally dying at week 49. Animals at 500 mg/kg bw per day had increased absolute weights of the testes, heart, and kidney and an increased amount of cancellous bone in the metaphysis. Ten animals at 250 mg/kg bw per day showed gross pituitary lesions, and one male and two females had malignant pituitary tumours. The NOEL was 50 mg/kg bw per day (Webb & Hansen, 1963).

Fifty weanling albino rats equally divided by sex were maintained on a dry diet containing methyl salicylate at concentrations calculated to result in an average daily intake of 0, 35, or 100 mg/kg bw for 2 years. Regular evaluation of growth, survival, food consumption, general physical condition, blood and urine parameters, and necropsy revealed no adverse effects associated with administration of the test substance. The NOEL was 100 mg/kg bw per day (Packman et al., 1961). However, these results were reported only in an abstract.

Groups of two male and two female beagles were given a capsule containing methyl salicylate at a dose of 0, 50, 150, or 350 mg/kg bw per day on 6 days/week for 2 years. The animals were weighed weekly, and haematological end-points were measured three times before initiation of the experiment and at 2 weeks, 1, 3, and 6 months, 1 year, and 2 years. Microscopy was performed on the three surviving dogs at the highest dose. Organs were weighed at necropsy, and selected tissues were examined histologically. Apart from two animals that died of unrelated diseases, no deaths occurred during the study, and all of the haematological values were normal. The groups at the two higher doses had enlarged livers and enlarged hepatic cells. Animals at 150 mg/kg bw per day showed growth retardation. The NOEL was 50 mg/kg bw per day (Webb & Hansen, 1963).

The results of short- and long-term studies are summarized in Table 4.

Studies of the genotoxicity of these compounds are summarized in Table 5.