The structurally related mycotoxins beauvericin and enniatins are produced mainly by Fusarium fungi that invade and grow on crops, and may produce them under moist and cool conditions. They are predominantly found in cereal grains and their products. The European Commission (EC) has asked the European Food Safety Authority (EFSA) for a scientific opinion on the risks to human and animal health related to the presence of beauvericin and enniatins in food and feed. The opinion should include an evaluation of the toxicity of beauvericin and enniatins for humans, considering relevant toxicological endpoints and toxicological relevance of beauvericin and enniatins in food; an exposure assessment of the European Union population to beauvericin and enniatins; and an assessment of the co-occurrence of beauvericin and enniatins with other Fusarium toxins. With regard to the animal risk assessment, the opinion should include the determination of exposure to beauvericin and enniatins for different animal species. It should also consider carry-over of beauvericin and enniatins from the feed to the food products of animal origin. The EC request also asked for identification of the feed materials that could be considered as sources of beauvericin and enniatins and the characterisation of the distribution of levels of contamination in feed materials, and an assessment in feed of the co-occurrence of beauvericin and enniatins with other Fusarium toxins.
Beauvericin is a cyclic hexadepsipeptide that consists of alternating D-α-hydroxy-isovaleryl-(2-hydroxy-3-methylbutanoic acid) and N-methylphenylalanine moieties. Enniatins are structurally related mycotoxins representing a large group of cyclic hexadepsipeptides. The chemical structure of enniatins comprises D-α-hydroxy-isovaleryl-(2-hydroxy-3-methylbutanoic acid) and N-methylamino acid residues. In enniatins of type A and B, the amino acid units are aliphatic N-methyl-valine or N-methyl-isoleucine or the mixtures of these two. The subunits are linked with peptide bonds and intra-molecular esters (lactone) bonds, forming a cyclic depsipeptide. So far, 29 naturally occurring enniatin analogues have been identified but only seven enniatins (enniatins A, A1, B, B1, B2, B3 and B4) have been found in cereals. Four of these, namely enniatins A, A1, B and B1, have been most frequently detected in foods and feeds.
No previous exposure or risk assessments on beauvericin and enniatins were identified in Europe or in other regions of the world. However, the EFSA Panel on Contaminants in the Food Chain (CONTAM) noted that an antibiotic and anti-inflammatory nasal/oromucosal spray agent containing enniatins called fusafungine has been used for the treatment of upper respiratory infection.
Methods for the determination of beauvericin and enniatins are established and can be applied for the analysis of food and feed commodities including animal products. Quantification of beauvericin and enniatins is carried out mostly by liquid chromatography coupled with (multi-stage) mass spectrometry (LC-MS(/MS)) often within a multi-analyte approach without any clean-up. For accurate quantification, tandem MS (LC-MS/MS) is considered superior to LC-UV, since it can better address matrix effects. No other analytical approaches, such as, for example immunochemical methods, are available for the determination of these toxins. None of the applied methods has been formally evaluated in inter-laboratory validation studies and there are no certified reference materials available for beauvericin and enniatins.
Following a call for data by EFSA in October 2010, 14 data providers submitted data coming from 12 European countries, and these covered food and feed, but also unprocessed grains of undefined end-use (also referred to as unprocessed grains in this scientific opinion). Thus, based on product category information, separate datasets were extracted and analysed for food, feed and unprocessed grains. Data were reported on samples collected between 2000 and 2013, with the majority of results collected after 2010. Occurrence data were obtained for beauvericin in food (n = 732), feed (n = 861) and unprocessed grains (n = 554), and for enniatins in food (n = 4 251), feed (n = 3 640) and unprocessed grains (n = 2 647). From these 12 685 analytical results, 4 823 results on food, 4 501 results on feed and 3 196 results on unprocessed grains fulfilled the quality criteria applied and have been used in the assessment. The proportions of results below the limit of detection (LOD) or limit of quantification (LOQ) were 80 % for beauvericin and 63 % for enniatins in food, 79 % for beauvericin and 32 % for enniatins in feed, and 46 % for beauvericin and 24 % for enniatins in unprocessed grains.
For enniatins, the CONTAM Panel decided to consider the sum of the concentrations of four enniatins, i.e. enniatins A, A1, B and B1, for the assessment since (i) these are the most frequently detected enniatins in foods and feeds, (ii) the toxicity database contained almost no data on individual enniatins but did contain some data on enniatin mixtures mainly containing these four toxins, and (iii) these enniatins are structurally similar and have similar modes of action. These considerations support the view that, as a first approximation, their effects might be comparable.
For food, the highest mean concentrations of beauvericin were measured in ‘Dried fruits’, followed by ‘Oilseeds’ and ‘Cereal-based food for infants and young children’. The highest mean concentrations for the sum of enniatins were in ‘Coffee beans’, followed by ‘Pasta (raw)’. For feed and unprocessed grains, the highest mean concentrations of beauvericin were measured in ‘Maize gluten’ and for the sum of enniatins in ‘Rye grain’ and ‘Barley’. Physical processing such as cleaning, sorting and milling results in a reduction of the concentrations of beauvericin and enniatins in the refined product and an increase in the cereal by-products. Beauvericin and enniatins are substantially stable during commercial cereal processing, including hot drying and ensiling procedures.
The occurrence data submitted to EFSA indicated a high relationship between each of the four enniatins and between the occurrence of beauvericin and enniatins in cereal grains. This relationship is expected because beauvericin and enniatins are structurally closely related and are produced by Fusarium species. The limited number of data submitted to EFSA did not allow any detailed analysis of the co-occurrence of beauvericin and enniatins with other Fusarium toxins. However, it seems that in cereal grains beauvericin and enniatins co-occur regularly with other Fusarium toxins, e.g. deoxynivalenol, and in particular moniliformin and fumonisins.
The CONTAM Panel estimated total chronic and acute dietary exposures to beauvericin and to the sum of enniatins. For humans, the chronic exposure was calculated across 17 European countries, using lower bound (LB) and upper bound (UB) mean concentrations and individual consumption data for different age groups. The acute exposure was calculated across 22 European countries, using the highest reliable percentile of concentrations, and individual consumption data for different age groups. Chronic exposure to beauvericin (minimum LB to maximum UB) ranged from 0.003 μg/kg b.w. per day to 0.050 μg/kg b.w. per day for the mean exposure, and from 0.006 to 0.093 μg/kg b.w. per day for the 95th percentile exposure. The highest acute exposure estimates for beauvericin were 0.05 μg/kg b.w. per day (mean) and 0.10 μg/kg b.w. per day (95th percentile). The mean chronic exposure to the sum of enniatins ranged from 0.42 to 1.82 μg/kg b.w. per day and the 95th percentile exposure ranged from 0.91 to 3.28 μg/kg b.w. per day. The highest acute exposure estimates of the sum of enniatins were 4.67 μg/kg b.w. per day (mean) and 10.1 μg/kg b.w. per day (95th percentile). Toddlers were in general the age group with the highest dietary chronic and acute exposure to both beauvericin and the sum of enniatins. ‘Grains and grain-based products’ made the largest contribution to the beauvericin and enniatins exposure. Important contributors were ‘Bread and rolls’, ‘Pasta (raw)’ and ‘Fine bakery wares’ in all age groups, and ‘Coffee (beverage)’ only for the sum of enniatins in the adults population group. In infants and toddlers, the ‘Cereal-based foods for infants and young children’ also made an important contribution to the total exposure. The data on vegetarians were limited and do not indicate a marked difference from the general population in the dietary exposure to beauvericin and the sum of enniatins.
Acute and chronic exposure of livestock to beauvericin and the sum of enniatins (enniatins A, A1, B and B1) were calculated for the different animal species using the mean and 95th percentile LB and UB occurrence data in cereal grains, their products and by-products. Exposure to beauvericin and enniatins by livestock is primarily from consuming cereal grains and cereal by-products. Because almost half of the data provided to EFSA were in categories ‘Unprocessed grains’ and ‘Grains as crops’, rather than defined by grain type, all the data relating to the data on specific cereal grains were aggregated with the data in more general categories of ‘Unprocessed grains’ and ‘Grains as crops’. Although beauvericin and enniatins have also been reported in other feed materials, these are normally at markedly lower concentrations, and therefore exposure assessments were restricted to those livestock for which cereal grains and/or their by-products represented at least 30 % of feed consumed. As a result, the animal species considered were dairy cows, cereal-fed beef cattle, piglets and pigs, laying hens, broiler chickens, turkeys, ducks, rabbits, dogs, cats and horses. The highest chronic beauvericin UB exposure was 0.86 μg/kg b.w. per day for ruminants and horses, and for the sum of enniatins the highest UB chronic exposure was 27.8 μg/kg b.w. per day for poultry. The highest acute beauvericin exposure was 1.89 μg/kg b.w. per day for poultry, and for the sum of enniatins the highest acute exposure was also for poultry at 113 μg/kg b.w. per day.
The available data on the toxicokinetics of beauvericin and enniatins are limited. In vitro data indicate that beauvericin and enniatins are absorbed and rapidly metabolised to a range of uncharacterised metabolites. For enniatins, the ratio between the different metabolites formed is species dependent. The oral bioavailability has not been determined in experimental animals, but both beauvericin and enniatins are found in rat plasma after oral exposure. During the exposure period, the plasma concentrations of enniatin A increased in rats given high concentrations in the feed for four weeks. The oral absorption of enniatin B in pigs after intragastric dosing was 91 % and a large difference in plasma concentration of the different enniatins in a mixture of four enniatins was observed in one pig. The elimination was rapid, but metabolites were not analysed. Beauvericin and enniatins were detected in eggs of laying hens, and in turkey and broiler tissues. Some metabolites of enniatin B were identified in serum, liver and eggs. The carry-over rate of beauvericin, enniatin B and enniatin B1 from feed to broiler meat, liver and skin and to laying hen eggs is low. This finding indicates that residues of beauvericin and enniatins from poultry may only marginally contribute to the human exposure.
Beauvericin and enniatins are cytotoxic, and the effects seem to be related to ionophoric properties. For beauvericin, the LD50 for acute toxicity was 100 mg/kg b.w. upon oral administration to mice. An LD50 of 350 mg/kg b.w. was reported for fusafungine (a mixture of enniatins) in mice upon oral administration. A sub-acute toxicity study on enniatin A with some limitations showed no adverse effects. There are no other in vivo toxicological studies available on sub-chronic, chronic, reproduction and developmental toxicity, neurotoxicity or carcinogenicity neither for beauvericin nor for enniatins. In vitro genotoxicity data are equivocal for beauvericin but some studies suggested a potential genotoxic effect. In vitro genotoxicity data for enniatins were negative. The CONTAM Panel noted some limitations in the study designs. There are no in vivo genotoxicity data for either beauvericin or enniatins. In vitro studies indicated immunotoxicity and haematotoxicity/myelotoxicity of beauvericin and enniatins. One in vivo study of enniatin A was available, but did not allow conclusions on functional implications. There are no reports on adverse effects in humans caused by food contaminated with beauvericin and enniatins.
For farm and companion animals the only available studies identified on adverse effects following oral administration were for beauvericin in poultry. One study was performed by using purified beauvericin administered by intubation in ducklings for seven days. No median lethal dose response at doses up to 100 mg/kg b.w. was observed. All the other studies were on broilers, laying hens and turkeys by using diets naturally contaminated with a combination of Fusarium toxins. The results showed that broilers and laying hens were not sensitive to different Fusarium toxin combinations at the tested levels up to 12.72 and 11.23 mg enniatin B per kg feed, and up to 4.06 and 3.60 mg enniatin B1 per kg feed, respectively. The results showed that broiler chickens, laying hens and turkeys were not sensitive to different Fusarium toxin combinations at the tested levels of up to 12.6, 8.93 mg and 2.48 mg beauvericin per kg feed, respectively. The CONTAM Panel converted the concentration of beauvericin and enniatins causing no adverse effects in broilers, laying hens and turkeys to respective no-observed-adverse-effect levels (NOAELs). For beauvericin, the CONTAM Panel identified NOAELs of 1 220, 536 and 136 µg/kg b.w. per day for broiler chickens, laying hens and turkeys, respectively. The CONTAM Panel identified NOAELs of 763 and 244 µg/kg b.w. per day for broiler chickens, and of 674 and 216 µg/kg b.w. per day for laying hens for enniatin B and enniatin B1, respectively.
The CONTAM Panel noted that there were insufficient data to establish a tolerable daily intake (TDI) or/and an acute reference dose (ARfD) for beauvericin or the sum of enniatins. The Panel also noted that a risk assessment was not possible for dietary exposure to beauvericin and enniatins given the overall lack of toxicity data and the limitations of the data available for acute toxicity and genotoxicity in experimental animals.
Considering the fact that there are only few or no relevant toxicity data available for beauvericin and the four enniatins A, A1, B, B1, and that the chemical structure of these five mycotoxins are known, the CONTAM Panel explored also the concept of the threshold of toxicological concern (TTC) and concluded that a risk assessment would require compound-specific toxicity data.
Given the overall lack of toxicity data, the CONTAM Panel decided to investigate the following two strategies to obtain some insight in the possible risks of beauvericin and the sum of enniatins at the estimated levels of exposure:
(1) to compare the estimated acute exposure levels with the LD50 values for beauvericin and the drug fusafungine (a mixture of enniatins);
(2) to compare the estimated chronic exposure levels with the doses reported to cause adverse effects upon therapeutic use of the drug fusafungine taking a worst-case approximation for converting the nasal/oromucosal dose levels to oral dose levels.
For strategy (1), the CONTAM Panel calculated the margins between the estimated acute dietary exposure in humans and the LD50 values for beauvericin and fusafungine. For beauvericin, the margins ranged from about 10 × 106 to 2 × 106 for the mean dietary exposure, and from about 5 × 106 to 1 × 106 for the 95th percentile dietary exposure. Comparison of the exposure values to the LD50 value for fusafungine revealed margins that ranged from about 35 × 106 to 7 × 106 for the mean dietary exposure, and from about 17.5 × 106 to 3.5 × 106 for the 95th percentile exposure. For the sum of enniatins, the margins ranged from about 99 000 to 21 500 for the mean dietary exposure, and from about 45 500 to 9 900 the 95th percentile dietary exposure, when using the LD50 of fusafungine.
For strategy (2), the Panel calculated the margins between the estimated chronic dietary exposure in humans and the estimated LOAEL of fusafungine for the sum of enniatins and, in the absence of toxicity data on repeated exposure, also for beauvericin. For beauvericin, the resulting margins ranged from about 57 000 to 1 800 for the mean dietary exposure, and from about 17 000 to 1 000 for the 95th percentile dietary exposure. For the sum of enniatins, the margins ranged from about 400 to 50 for the mean dietary exposure, and from about 190 to 30 for the 95th percentile dietary exposure. The CONTAM Panel noted the widespread use of fusafungine based drugs and their over-the-counter availability as well as unsystematic reporting and collection of information available on the adverse effects. Therefore, the LOAEL derived from the information on adverse effects of fusafungine applied to humans as a drug may be highly uncertain, and when calculated from the prescribed dosing it could lead to an over- or underestimation of a LOAEL.
The CONTAM Panel concluded that the large margins obtained for acute exposure to beauvericin and enniatins do not indicate concern for human health. There might be a concern with respect to chronic exposure but no firm conclusion could be drawn.
For animal health risk characterisation, the acute and chronic exposure levels of beauvericin and enniatins (sum of enniatins A, A1, B and B1) for the different animal species were estimated using the LB and UB occurrence data for beauvericin and the sum of enniatins in cereal grains and their products. Although beauvericin and enniatins are known to be present in other feed materials, there were insufficient data on concentrations in these feeds to include them in the exposure assessments.
The estimated chronic exposure for poultry at the currently reported concentrations in feed indicate that adverse health effects of feed containing beauvericin and enniatins are unlikely. For other livestock species and companion animals, the lack of LOAELs/NOAELs precludes the assessment of chronic health risks associated with beauvericin and the sum of enniatins in feed. In the absence of adequate acute toxicological data, the CONTAM Panel used alternative approaches for animal health risk characterisation for beauvericin and the sum of enniatins using LD50 values derived from studies with beauvericin and fusafungine. The CONTAM Panel concluded that acute adverse health effects of feed containing beauvericin and enniatins are unlikely for the livestock species and companion animals under current feeding practices.
The most important contributions to the uncertainty in this evaluation originated from the lack of toxicity data both for humans and experimental animals for beauvericin and enniatins. Additional sources of uncertainty were a LOAEL derived from the data on the use of the drug fusafungine, and the alternative approaches used to inform on possible risks for humans. No toxicological data were available on a number of farm animals and companion animals. Additional uncertainties arose by using the LD50 of the human drug fusafungine for animals. Furthermore, the use of the sum of only four enniatins (A, A1, B and B1) for human and animal exposure, where no information on possible combined effects of the single enniatins was available, or effects of enniatins other than these four, added uncertainty. The CONTAM Panel considered the uncertainties resulting from the lack of toxicity data on beauvericin and enniatins prevent a risk assessment for humans and most livestock.
To further improve the risk assessment on beauvericin and enniatins in food and feed, the CONTAM Panel recommends the use of LC-MS/MS-based methods, since matrix effects can be better addressed and are able to quantify concentrations below 1 µg/kg. The analytical methods should be evaluated in inter-laboratory validation studies, and performance criteria should be developed. (Certified) reference materials should be developed for the determination of beauvericin and enniatins in food and feed. Further investigations on the fate of beauvericin and enniatins during the preparation of grain-based products (especially bread and rolls, fine bakery wares and pasta) should be conducted. To perform a human risk assessment, in vivo toxicity data on beauvericin and enniatins are needed. A study investigating possible health effects likely to arise from repeated exposure (i.e. 90-day study), including effects on the nervous, immune and endocrine systems, as well as screening of possible effects on reproduction and development, is required. Additional in vitro and in vivo genotoxicity data are needed. Depending on the outcome of these studies, additional studies might be required. The CONTAM Panel recommends that, in the absence of toxicological data for most livestock and companion animals, studies on adverse effects are needed to identify NOAELs. Given their relatively high susceptibility to Fusarium toxins and ionophores, respectively, pigs and horses should be given the highest priority. Lastly, further data on the co-occurrence of beauvericin and enniatins with other Fusarium toxins in food and feed, and the possible combined effects are required.