(a) Occurrence in milk through ammoniated forage

Exposure can occur from the consumption of foods contaminated with 4-methylimidazole, which is formed as a result of the interaction of ammonia with reducing sugars. Forage — typically hay and straw — is sometimes treated with anhydrous ammonia to improve its quality (e.g. increase the non-protein nitrogen content) and digestibility (Waagepetersen & Vestergaard, 1977). Imidazoles (such as 4-methylimidazole) and pyrazines appear to be the dominant groups of toxic by-products formed from the interaction of ammonia with reducing sugars. Experimental studies have shown that higher concentrations of sugar and ammonia, higher temperatures, higher water activity and longer reaction times increased the amount of 4-methylimidazole (formed at the pH achieved by the addition of ammonia) (Bergström, 1991). Perdok & Leng (1987) reported that 4-methylimidazole was not present in most types of untreated roughage, but was found at concentrations ranging from 8 to 43 mg/kg in thermo-ammoniated roughage. Average concentrations of 0.72 μg/mL 4-methylimidazole were detected in the plasma of sheep fed ammoniated tall fescue that contained an average concentration of 64.36 mg/kg of the chemical (Karangwa et al., 1990a). 4-Methylimidazole has been identified in the plasma, urine and milk of cows and sheep fed ammoniated forage (Müller et al., 1998a, b). Müller et al., (1998a) reported that the concentrations of 4-methylimidazole in an ammoniated forage-fed (90 μg/g dry matter) ewe were 0.07 μg/mL in plasma, 0.23–0.31 μg/mL in milk and 21 μg/mL in urine. The plasma concentration in one of the ewes' suckling lambs that developed toxicosis was 0.01 μg/mL. Similar concentrations were found in the plasma and milk of ewes fed ammoniated seed hay (with 4-methylimidazole concentrations greater than 100 μg/g dry matter) for 7 days (Sivertsen et al., 1993). In a dairy cow fed ammoniated forage containing 4-methylimidazole (58 μg/g dry matter), concentrations of the chemical in plasma, milk and urine were 0.28, 2.7 and 5.8 μg/mL, respectively (Müller et al., 1998a).

(b) Occurrence in food and drinks containing caramel colourings

4-Methylimidazole is found in ammonia and ammonia-sulfite process caramel colourings. Caramel colourings are produced by heating carbohydrates with specified reagents under defined temperatures and pressures, which results in a dark brown colouring with a characteristic odour of burnt sugar. Their use accounts for 95% by weight of the permitted colour additives used in food. They have been classified by the Joint FAO/WHO Expert Committee on Food Additives (JECFA), and the European Union Scientific Committee for Food into four classes, two of which are prepared using compounds that contain ammonia (reviewed by Chappel & Howell, 1992; Houben & Penninks, 1994). Class III ammonia caramels are commonly used in various bakery products, soya-bean sauces, brown sauces, gravies, soup aromas, brown (dehydrated) soups, brown malt caramel blends for various applications, vinegars and beers, especially in certain dark-brown beers. Their use accounts for 20–25% of the total use of caramel colourings in the USA and for about 60% in Europe. Class IV ammonia-sulfite caramels are used in soft drinks, pet foods and soups (Houben & Penninks, 1994), and account for approximately 70% of the caramel colourings produced worldwide (Licht et al., 1992a). Reported concentrations of 4-methylimidazole in caramel colourings are provided in Table 1.1. Licht et al., (1992b) reported 4-methylimidazole concentrations ranging from < 5 to 184 mg/kg in 40 commercial Class III caramel colourings that met JECFA guidelines for 2-acetyl-4-tetrahydroxybutylimidazole. Other studies have reported higher concentrations in some samples, ranging up to 463 mg/kg. In general, higher 4-methylimidazole levels have been found in Class IV caramel colourings; a study of 90 commercial products found levels ranging from 112 to 1276 mg/kg (see Table 1.1).

Table 1.1

Concentrations of 4-methylimidazole in caramel colourings and food and beverages.

Long-term dietary exposure to caramels among children aged 1–10 years has been estimated based on analytical data in 11 European countries (EFSA, 2010). Median exposure ranged from 4.3 to 41 mg/kg body weight (bw) per day for ammonia-sulfite caramels (class IV) and from 32 to 105 mg/kg bw per day for ammonia caramels (class III).

Reported concentrations of 4-methylimidazole ranged from 1.58 to 28.03 mg/kg in dark beer (Klejdus et al., 2006), from 0.3 to 1.45 mg/kg in coffee (Casal et al., 2002; Lojková et al., 2006) and from 0.30 to 0.36 μg/mL in soda (Moon & Shibamoto, 2010). Yoshikawa & Fujiwara (1981) measured 4-methylimidazole in various foods and beverages (see Table 1.1). The highest levels were found in Worcestershire sauce (up to 3.4 mg/kg) or foods cooked in soya sauce (up to 3.2 mg/kg).

(a) Absorption, distribution and excretion

Previous studies have shown species differences in the disposition of 4-methylimidazole.

In rats, 5 minutes after a single intraperitoneal injection of 216 mg/kg bw, the uptake of 4-methylimidazole was highest in the intestines, followed by the liver, blood, stomach and kidney (Hidaka, 1976). The compound was excreted unchanged in the urine, beginning approximately 30 minutes after injection, and reached approximately 90% within 8 hours (Hidaka, 1976).

In ewes, the absorption and elimination of a single oral dose of 4-methylimidazole followed frst-order kinetics. One half of an oral dose (20 mg/kg bw) of 4-methylimidazole was absorbed within about 27 minutes, and the maximum plasma level was reached 5 hours after administration (Karangwa et al., 1990b). Bioavailability calculated from plasma data from three ewes was 69%, and the biological half-life was 9.37 hours. Only 0.07 mg/kg of the dose was recovered in the urine as the unchanged parent compound. Metabolites of 4-methylimidazole were not detected by high-performance liquid chromatography. [The Working Group noted that the sensitivity of this assay was difficult to evaluate.]

In goats and heifers, the mean residence time of 4-methylimidazole administered orally or intravenously was about 5 hours, and the volume of distribution was 0.9 L/kg bw. 4-Methylimidazole and its metabolites were excreted mainly in the urine, but also in the milk and faeces, and the administered dose was distributed mainly in the liver, kidney and lung (Nielsen et al., 1993). 4-Methylimidazole was found in the milk following its oral administration to pregnant and postpartum cows (Morgan & Edwards, 1986).

Following oral administration by gavage of 5, 50 or 150 mg/kg bw 4-methylimidazole (¹⁴C-radiolabelled) to F344/N rats, peak plasma concentrations were reached at 0.5, 1.0 and 3.0 hours, respectively (Yuan & Burka, 1995). At 150 mg/kg, the plasma concentration of [¹⁴C]4-methylimidazole was almost constant during the first 5 hours; at lower doses, the decline was more rapid. The estimated terminal half-life was dose-dependent. The authors suggested that the elimination of parent 4-methylimidazole was saturable. From the total urinary recovery of parent 4-methylimidazole, the estimated bioavailability was approximately 60–70%. Little or no metabolism of 4-methylimidazole was found. Only one minor hydrophilic metabolite was present in the urine and plasma. Faecal, biliary and respiratory elimination of radio-activity were negligible.

(b) Metabolism

Four metabolites were determined in the urine of goats and heifers given 4-methylimidazole (Nielsen et al., 1993), three of which were identified as 5-methyl-hydantoin, 2-methyl-hydantoic acid and urea. The high polarity of the fourth product prevented further characterization.

(c) Toxicokinetic models

After a single oral administration by gavage of 4-methylimidazole (10, 50 or 100 mg/kg bw) to male and female F344/N rats, the plasma concentration versus time data could be described by a one-compartment model, with no lag phase, and frst-order absorption and elimination for both males and females based on the findings of Yuan & Burka (1995) (NTP, 2007). The absorption half-life ranged from 5 to 23 minutes and decreased with dose. The elimination half-life ranged from 1 to 8 hours and increased with dose. The plasma concentration versus time data following intravenous administration of 10 mg/kg bw 4-methylimidazole was described as a one-compartment model with frst-order elimination. From comparisons of the area under the concentration versus time curves for the two routes of administration, bioavailability was determined to be greater than 85%.

(a) Mutations

4-Methylimidazole (up to 10 000 μg/plate) was not mutagenic in Salmonella typhimurium strains TA97, TA98, TA100 or TA1535 when tested in the presence or absence of 10% or 30% hamster or rat liver metabolic activation systems (detailed protocol presented by Zeiger et al., 1988; NTP, 2007). Class III and IV caramel colourings contain various concentrations of 4-methylimidazole, and did not to induce mutagenic activity in S. typhimurium strains TA98, TA100, TA1535, TA1537 or TA1538. In these studies, concentrations of 4-methylimidazole in class III preparations were 9–463 mg/kg, and those in class IV were 146–215 mg/kg (Allen et al., 1992). Class III and IV caramel colourings were also negative in the S. typhimurium Ames test and Saccharomyces cerevisiae gene conversion assays. In these studies, concentrations of 4-methylimidazole in class III preparations were 34–463 mg/kg, and those in class IV were 107–387 mg/kg (Brusick et al., 1992).

(b) Chromosomal effects

No increases in the frequency of micro-nucleated erythrocytes were observed in the bone marrow of male rats or male mice (detailed protocol presented by Shelby et al., 1993; NTP, 2007) administered three intraperitoneal injections of 4-methylimidazole at 24-hours intervals or in peripheral blood samples from male and female mice fed the compound in the diet for 14 weeks (detailed protocol presented by MacGregor et al., 1990; NTP, 2007). No signifcant alterations in the percentage of polychromatic erythrocytes, an approximate indicator of bone marrow toxicity, were seen in the bone marrow or peripheral blood of mice; however, the percentage declined with increasing dose of 4-methylimidazole and was significantly depressed at the highest dose in the bone marrow of male rats (NTP, 2007).

Class III caramel colouring did not induce chromosomal damage in Chinese hamster ovary cells (Allen et al., 1992). In the study of Brusick et al., (1992), class IV caramel colouring gave negative results in the chromosomal aberration assay while class III colouring was weakly clastogenic only in the absence of metabolic activation or in the presence of heat-inactivated metabolic activation. Class IV caramel colouring was not clastogenic in Chinese hamster ovary cells in vitro in either the presence or absence of metabolic activation, whereas the weak clastogenic effect of class III caramel colouring was abolished in the presence of metabolic activation. Moreover, in vivo, class III caramel colouring administered orally to mice did not induce micronuclei in the bone marrow.