Following a request from the European Commission, the Panel on Contaminants in the Food Chain (CONTAM Panel) was asked to deliver a scientific opinion on the risks for human health related to the presence of 3- and 2-monochloropropanediol (MCPD), and their fatty acid esters, and glycidyl fatty acid esters (GE) in food.
3-Monochloropropane-1,2-diol (3-MCPD) and 2-monochloropropane-1,3-diol (2-MCPD) are chlorinated derivatives of glycerol (1,2,3-propanetriol). 3- and 2-MCPD and their fatty acid esters are among non-volatile chloropropanols, identified in the late 1970s in the composition of hydrolysed vegetable protein (HVP) which is used as a savoury flavour-enhancing food ingredient. 3- and 2-MCPD fatty acid esters are produced in vegetable oils on refining, and they contain the fatty acids common to the parent oils and fats in a similar ratio.
Glycidol is associated with the formation and decomposition of 3- and 2-MCPD. It forms monoesters with fatty acids (GE) during the refining of vegetable oils.
Chloropropanols are formed in HVP during the hydrochloric acid-mediated hydrolysis step of the manufacturing process. In food production, chloropropanols form from the reaction of endogenous or added chloride with glycerol or acylglycerols, although the precise mechanism has not yet been elucidated. The major proposed routes of formation of 3- and 2-MCPD fatty acid esters include a direct nucleophilic attack by chloride ions on the acylglycerol carbon atom attached to an ester or hydroxyl group; the formation via chlorination of a GE; and formation via a cyclic acyloxonium ion or a cyclic acyloxonium free radical.
The only identified source of GE in food is refined vegetable oil where they appear to be formed during the heating of diacylglycerol (DAG) or monoacylglycerol under the high temperature conditions of deodorisation.
GE formation is believed to be independent from MCPD fatty acid ester formation, although they may also be formed by elimination of hydrochloric acid from MCPD monoesters that have a vicinal chlorohydrin structure. These monoesters are 3-MCPD that is esterified in the 1- or 2-position and 2-MCPD monoesters.
Processing conditions used in the manufacture of HVP produce free 3- and 2-MCPD. Steps are now taken routinely to both reduce their formation and to lower their levels, but the problem is not eliminated. Glycidyl esters are formed from DAG on heating vegetable oils to temperatures in excess of 200°C, for example during the deodorisation stages of refining, and are therefore a particular problem in palm oil, which can have a high (4–12%) DAG content.
Analytical methods for free 3- and 2-MCPD in foods are well characterised, validated for a suitable range of foods and fit for purpose. There are no suitable methods for unstable free glycidol. Indirect methods for ester-bound 3- and 2-MCPD and glycidol in foods are well characterised for the important range of foods. Three validated American Oil Chemists’ Society (AOCS) methods exist and provide directly comparable results. The methods for free MCPD do not include MCPD released from esters, and the methods for ester-bound 3- and 2-MCPD and glycidol do not provide data for the free compounds. Two methods were therefore applied – one for free MCPD and one for MCPD and glycidol released from bound forms.
In its exposure assessment, the CONTAM Panel considered a total of 7,175 occurrence data on 3-MCPD, 2-MCPD and glycidol (collated as measures were implemented to reduce the levels of these compounds in edible fats/oils). Data on glycidol were only available from the ester-bound form. Three categories of analytical data were considered – one on 3-MCPD (in free form) in soy sauce, HVP and related products (702 data points); another on 3- and 2-MCPD from esters and glycidol from esters in oils/fats (4,754 data points); and a third one on 3- and 2-MCPD (free and from esters) and glycidol (from esters) in food groups other than those mentioned above (1,719 data points). In the third category, in most cases, the contribution to the total 3- and 2-MCPD from the free form was included, while the results on glycidol were only from esters. In fats and oils, only the ester-bound forms were analysed because the contribution of the free forms was considered negligible. More than half of the data referred to fats and oils, but other food groups where the presence of these substances is expected were also represented in the data set. Where possible, for food groups not represented in the data set, the occurrence of 3-, 2-MCPD and glycidol was calculated using a model based on the available data on fats and oils, taking into account the market share of the different oils in Europe. This model was also used to estimate the contribution of 3- and 2-MCPD from ester-bound forms in food groups for which data were available only for the free forms.
The highest occurrence values were found in the food group ‘Fats and oils’, with ‘Palm oil/fat’ showing a mean middle bound (MB) level of 2,912 µg/kg for 3-MCPD (from esters), 1,565 µg/kg for 2-MCPD (from esters) and 3,955 µg/kg for glycidol (from esters). Lower mean MB levels were calculated for other oils, ranging between 48 and 608 µg/kg for 3-MCPD (from esters), between 86 and 270 µg/kg for 2-MCPD (from esters) and between 15 and 650 µg/kg for glycidol (from esters). ‘Margarine, normal fat’ had mean MB levels of 668 µg/kg for 3-MCPD (from esters), 236 µg/kg for 2-MCPD (from esters) and 582 µg/kg for glycidol (from esters). Among food groups other than fats and oils, the highest levels were observed in ‘Potato crisps’ (mean MB levels of 216 µg/kg for total 3-MCPD, 135 µg/kg for total 2-MCPD and 110 µg/kg for glycidol from esters), ‘Hot surface cooked pastries’ (mean MB levels of 247 µg/kg for total 3-MCPD, 123 µg/kg for total 2-MCPD and 137 µg/kg for glycidol from esters), ‘Cookies’ (mean MB levels of 200 µg/kg for total 3-MCPD, 103 µg/kg for total 2-MCPD and 134 µg/kg for glycidol from esters) and ‘Short crusts’ (mean MB levels of 154 µg/kg for total 3-MCPD, 79 µg/kg for total 2-MCPD and 149 µg/kg for glycidol from esters).
The exposure assessment for 3- and 2-MCPD was based upon the level of exposure to the parent compounds, regardless of their original form (i.e. as free or as ester of fatty acids), and referred to as 3-MCPD and 2-MCPD. Likewise, exposure to glycidol referred to the parent compound, although in this case, the original forms were exclusively as fatty acid esters.
Chronic dietary exposure to 3- and 2-MCPD and glycidol was assessed as mean and high (95th percentile, P95) exposure across dietary surveys. The exposure levels showed relatively little difference between lower bound (LB) and upper bound (UB) estimates, and the risk characterisation was therefore based on MB estimates of exposure. In all scenarios, the youngest population groups (‘Infants’, ‘Toddlers’ and ‘Other children’) showed the highest dietary exposure estimates.
The mean exposure to 3-MCPD was 0.5–1.5 µg/kg body weight (bw) per day (MB) across the dietary surveys for the age groups ‘Infants’, ‘Toddlers’ and ‘Other children’. The high exposure (P95) to 3-MCPD was 1.1–2.6 µg/kg bw per day (MB) across dietary surveys in these age groups. In adolescents and adult population groups (adults, elderly, very elderly), the mean exposure to 3-MCPD ranged from 0.2 to 0.7 µg/kg bw per day (MB) and the high exposure (P95) ranged from 0.3 to 1.3 µg/kg bw per day (MB).
The mean 2-MCPD exposure (MB) across dietary surveys ranged from 0.2 to 0.7 µg/kg bw per day, for ‘Infants’, ‘Toddlers’ and ‘Other children’. The high exposure (P95) to 2-MCPD was 0.5–1.2 µg/kg bw per day (MB) across dietary surveys in these age groups. In adolescents and adult population groups (adults, elderly, very elderly), the mean exposure to 2-MCPD ranged from 0.1 to 0.3 µg/kg bw per day (MB) and the high exposure (P95) ranged from 0.2 to 0.6 µg/kg bw per day (MB).
The mean exposure to glycidol was 0.3–0.9 µg/kg bw per day (MB) across the dietary surveys forthe age groups ‘Infants’, ‘Toddlers’ and ‘Other children’. The high exposure (P95) to glycidol was 0.8–2.1 µg/kg bw per day (MB) across dietary surveys in these age groups. In adolescents and adult population groups (adults, elderly, very elderly), the mean exposure to glycidol ranged from 0.1 to 0.5 µg/kg bw per day (MB). The high exposure (P95) in ‘Adolescents’ ranged from 0.4 to 1.1 µg/kg bw per day (MB) and in adults and older population groups ranged from 0.2 to 0.7 µg/kg bw per day (MB).
Exposure scenarios of infants receiving formula only, based on mean consumption and mean occurrence in the formula, resulted in daily intake of 2.4 µg/kg bw for 3-MCPD, 1.0 µg/kg bw for 2-MCPD and 1.9 µg/kg bw for glycidol. Using P95 occurrence data resulted in daily intake of 3.2 µg/kg bw for 3-MCPD, 1.6 µg/kg bw for 2-MCPD and 4.9 µg/kg bw for glycidol.
For ‘Infants’, the food groups ‘Infant and follow-on formulae’, ‘Vegetable fats and oils’ and ‘Cookies’ were the major contributors to 3- and 2-MCPD and glycidol exposure. For ‘Toddlers’, the food groups ‘Vegetable fats and oils’, ‘Cookies’ and ‘Pastries and cakes’ were the major contributors to 3- and 2-MCPD and glycidol exposure. ‘Infant formula’ and follow-on formula’ were also important contributors to 3- and 2-MCPD exposure. For ‘Other children’, the food groups with highest contribution to exposure to 3- and 2-MCPD and glycidol were ‘Pastries and cakes’, ‘Margarine and similar’ and ‘Cookies’. For glycidol, ‘Fried or roast meat’ was an additional relevant contributor. ‘Vegetable fats and oils’ also contributed to 3- and 2-MCPD, and glycidol exposure. For ‘Adolescents’, ‘Adults’, ‘Elderly’ and ‘Very elderly’, the major sources of 3- and 2-MCPD and glycidol were ‘Margarine and similar’ and ‘Pastries and cakes’. Additionally, ‘Fried or baked potato products’ were important contributors to 3- and 2-MCPD exposure while ‘Fried or roast meat’ and in some cases ‘Chocolate spreads and similar’ were important contributors to glycidol exposure.
3-MCPD and its dipalmitate fatty acid esters appear to be rapidly and efficiently absorbed following ingestion with extensive presystemic de-esterification occurring in the gastrointestinal tract of rats. Elimination of 3-MCPD from serum was also rapid following dosing with either the parent compound or its dipalmitate ester. 3-MCPD is extensively metabolised by routes including conjugation to glutathione and oxidation to b-chlorolactate and oxalic acid, with less than 5% appearing in the urine and faeces as parent compound. The majority of 3-MCPD is eliminated from serum within a few hours of dosing with either the parent compound or its dipalmitate ester.
No toxicokinetic data for 2-MCPD were identified. However, the difference in the structural localisation of the chorine within the molecule makes it unlikely that 2-MCPD exhibits the same metabolic pattern as 3-MCPD.
Glycidol and its fatty acid esters are efficiently absorbed following ingestion. Significant presystemic hydrolysis of GE occurs, although the de-esterification process appears to be more extensive in rats than in monkeys. Metabolism of the glycidol moiety proceeds rapidly by several enzymatic pathways, including glutathione conjugation and mercapturate formation. The glycidol moiety is predominantly excreted in urine as poorly described metabolites.
In short-term studies in rat, 3-MCPD produces severe renal toxicity at single intraperitoneal (i.p.) doses between 100 and 120 mg/kg bw, which persists for several weeks. Repeated oral doses also result in renal toxicity, and progressive nephropathy and renal tubule dilation can be seen after a daily dose as low as 5.2 mg/kg bw. The renal toxicity of 3-MCPD appears to reside with the R isomer.
3-MCPD administered to rats at 30 mg/kg bw per day impaired red blood cell function by decreasing haemoglobin content and inducing volume fraction changes consistent with normocytic and normochromic anaemia.
Neurotoxic effects such as hind limb paralysis were reported only at doses over 50 mg/kg bw per day following short-term exposure in mice.
In long-term studies, doses as low as 2 mg/kg bw per day 3-MCPD caused progressive nephrotoxicity (characterised by tubular hyperplasia), testicular toxicity (atrophy and arteritis) and mammary glandular hyperplasia in male rats and nephrotoxicity in female rats. Related to these effects, benign tumours of the testes (Leydig cell tumours), mammary gland (fibroadenoma) and kidney (tubular adenoma) developed.
Doses between 5 and 10 mg/kg bw per day 3-MCPD administered to rats can completely impair male fertility without changing sperm production. This effect has been demonstrated in several species including primates and is reversible. The no observed adverse effect level (NOAEL) of 3-MCPD on male fertility is not clear. Single and multiple doses of 3-MCPD administered to pregnant rats decreased the number of implantations and increased fetal loss but were not teratogenic. The NOAEL for multiple doses was 10 mg/kg bw per day for maternal toxicity and 30 mg/kg bw per day for fetal toxicity.
Despite some positive genotoxicity tests in vitro, there is no evidence that 3-MCPD is genotoxic in vivo in any organ tested, including the kidney and testis.
From the available information on 3-MCPD fatty acid esters, it can be concluded that the range of toxic effects for esterified 3-MCPD is the same as that seen for the free 3-MCPD, supporting the view that the esters are cleaved and toxicity is primarily exerted by 3-MPCD.
After equimolar multiple doses of 3-MCPD and 3-MCPD dipalmitate, the biochemical changes associated with renal toxicity are similar in pattern and magnitude. Both compounds produce an array of renal histopathology including glomerular lesions and tubular epithelial hyperplasia.
There is limited evidence that some esters of 3-MCPD have male antifertility effects at a similar molar dose to 3-MCPD, and degenerative changes in the spermatogenic tubules have been recorded after treatment with 3-MCPD fatty acid esters.
No studies on the in vitro genotoxicity of 3-MCPD fatty acid esters were identified. From the limited evidence (one study with different endpoints) available, there is no indication that 3-MCPD fatty acid esters are genotoxic in vivo.
The CONTAM Panel concluded that the kidney and testis appeared to be the main target organs for 3-MCPD-induced toxicity, the toxic effects being associated with oxidative metabolism of 3-MCPD to β-chlorolactaldehyde and β-chlorolactic acid. The inhibition of glycolysis by metabolites associated with the β-chlorolactate pathway was suggested as the possible nephron- and spermo-toxic mechanism of 3-MCPD.
The CONTAM Panel concluded that the Leydig cell and mammary gland tumours observed following long-term exposure to 3-MCPD were probably not relevant to humans.
The CONTAM Panel selected two long-term exposure studies where rats received 3-MCPD via drinking water to perform dose–response analysis for effects in the kidney and testis. The results of both studies were analysed and those showing a monotonic dose–response trend were selected for benchmark dose (BMD) analysis.
The CONTAM Panel established a tolerable daily intake (TDI) of 0.8 µg/kg bw per day for 3-MCPD. This was based on a chronic study in rats in which the lowest BMDL10 of 0.077 mg/kg bw per day for renal tubular hyperplasia in males was derived and application of an overall uncertainty factor of 100.
Noting the lack of specific data on 3-MCPD fatty acid esters and their hydrolysis, the CONTAM Panel confirmed that the toxicity of 3-MCPD fatty acid esters should be considered equivalent (on a molar basis) to that of the parent compound. Therefore, the CONTAM Panel concluded that the TDI of 0.8 µg/kg bw per day constitutes a group TDI for 3-MCPD and its fatty acid esters (expressed as MCPD equivalents).
The mean exposure to 3-MCPD was below the established group TDI of 0.8 µg/kg bw per day in ‘Adolescents’, ‘Adults’ and older age classes in all dietary surveys. In ‘Infants’, ‘Toddlers’ and ‘Other children’, half of the dietary surveys had mean exposure at or above the group TDI, up to a maximum of about 1.5 µg/kg bw per day in ‘Toddlers’ and ‘Other children’. The high exposure (P95) to 3-MCPD for ‘Infants’, ‘Toddlers’ and ‘Other children’ was above the group TDI in all dietary surveys, ranging between a minimum of 1.1 µg/kg bw per day in ‘Other children’ or roughly 1.5 µg/kg bw per day in ‘Infants’ and ‘Toddlers’ up to about 2.5 µg/kg bw per day in all the three age classes. The estimated exposure to 3-MCPD of infants receiving formula only was 2.4 µg/kg bw per day using mean occurrence and 3.2 µg/kg bw per day using P95 of occurrence; both values are above the group TDI, which is exceeded up to fourfold. The high exposure (P95) to 3-MCPD for adolescents was at or above the group TDI in half of the dietary surveys, with exposure estimates up to 1.4 µg/kg bw per day. For ‘Adults’ and the older age classes, only the maximum P95 of dietary exposure to 3-MCPD was around the group TDI.
There are limited data on the short-term toxicity of 2-MCPD. Acute median lethal dose (LD50) was estimated to be between 50 and 60 mg/kg bw in rats. A single i.p. dose of 200 mg/kg bw, although generally toxic, did not cause signs of renal toxicity. In a 28-day study in rats, daily doses of 16 or 30 mg/kg bw caused severe myopathy and nephrotoxicity. From 8 days of treatment, severe lesions leading to cell death developed in striated muscle, particularly in cardiac myocytes that resulted in heart failure and the death of some animals. These effects were not observed at 2 mg/kg bw per day. No data on long-term studies for 2-MCPD or 2-MCPD fatty acid esters were identified.
In vitro genotoxicity data on 2-MCPD were too limited to make any conclusion. No mammalian in vivo genotoxicity studies have been identified for 2-MCPD and 2-MCPD fatty acid esters.
2-MCPD did not induce kidney toxicity at doses at which 3-MCPD produced renal failure, enlarged kidneys and long-lasting diuresis. These differences were explained by the fact that metabolism of 2-MCPD to β-chlorolactaldehyde and β-chlorolactate cannot occur, which is believed to play an important role in nephrotoxicity of 3-MCPD. The underlying mechanisms for renal toxicity and the destruction of striated muscles, including the heart, are unknown.
Although the exposure data were available, it was not possible to undertake risk characterisation for 2-MCPD due to the lack of information.
For glycidol, neurotoxicity was observed after 28 days of treatment of rats with 200 mg/kg bw per day. Glycidol caused renal toxicity in repeated dose studies in rats and mice at doses in the range 150–400 mg/kg bw per day.
Two-year carcinogenicity studies in mice (25 and 50 mg/kg bw per day) and rats (37.5 and 75 mg/kg bw per day) showed induction of tumours in multiple organs from both sexes. Supporting evidence for carcinogenicity of glycidol was provided by a short-term study in a transgenic mouse strain.
Male anti-fertility effects have been noted in rats and mice. The lowest observed adverse effect level (LOAEL) was 25 mg/kg bw per day in the rat, resulting in a 36% reduction in epididymal sperm count. This may be attributed to conversion of glycidol to 3-MCPD in the stomach. Glycidol was maternally toxic in mice without producing any major external abnormalities in the fetus. Neurotoxicity was observed in male pups of rats exposed to a maternal dose of 49 mg glycidol/kg bw per day during pregnancy and weaning.
Glycidol and its esters, from which the free compound can be derived, possess a reactive epoxide moiety. There is strong evidence from in vitro data and some evidence from in vivo studies that glycidol is a genotoxic compound.
The CONTAM Panel only considered toxicity studies in animals with glycidol as no in vivo data were identified for glycidyl esters. Dose–response considerations were made for glycidol assuming a complete hydrolysis of the esters to free glycidol following ingestion. However, the dose–response data were considered inadequate for BMD modelling. Based on the EFSA Guidance on substances that are genotoxic and carcinogenic, T25 values were calculated for the incidence of tumours observed in rats and mice following long-term exposure to glycidol. A T25 of 10.2 mg/kg bw per day for peritoneal mesothelioma in male rats was used as the reference point.
In view of the genotoxic and carcinogenic potential of glycidol, a margin of exposure (MoE) approach was applied. MoE estimates were calculated by dividing the reference point of 10.2 mg/kg bw per day by the exposure levels. A MoE of 25,000 or higher was considered of low health concern.
For ‘Infants’, ‘Toddlers’ and ‘Other children’, the MoE estimates for the mean exposure ranged from 34,000 to 11,300; the MoE for high (P95) exposure ranged from 12,800 to 4,900. For ‘Adolescents’ and ‘Adults’, ‘Elderly’ and ‘Very elderly’ age groups, the MoE for the mean exposure ranged from 102,000 to 20,400, whereas at high (P95) exposure the range was from 51,000 to 9,300.
The MoE estimates corresponding to the P95 of exposure for ‘Infants’ were particularly low due to the contribution of glycidyl esters from infant formulae. The scenarios calculated for ‘Infants’ receiving only formula diet resulted in a MoE of about 5,400 for the mean occurrence and 2,100 for the P95 of occurrence.
In conclusion, estimated exposure substantially exceeding the group TDI for 3-MCPD is of concern; this is particularly seen in the younger age groups. Although there is a high uncertainty in the reference point used as a basis for the calculation of the MoEs for glycidol, the MoEs lower than 25,000 indicate a health concern.
The CONTAM Panel recommended to include all food groups potentially contaminated and foods where mitigation measures have been enforced in the future monitoring activities for 3-, 2-MCPD and glycidol. The enantiomeric composition of 3-MCPD and its fatty acid esters present in food should be studied. Further studies on the rates and degree of de-esterification and the metabolic fate for 3- and 2–MCPD fatty acid esters and GE were recommended. For 2-MCPD, the CONTAM Panel recommended generation of additional data to elucidate the long-term toxicity and the mode and mechanism of action of the substance. More extensive testing of the dose–response for carcinogenesis from chronic lifetime oral administration of glycidol and its esters in rats would reduce uncertainty in the risk assessment.