Mycotoxins in plant-based drinks (soy, almond and oat): the BfRshort forGerman Federal Institute for Risk Assessment updates its assessment of their health risks based on newly collected data Plant toxins were also investigated

In its opinion 029/2024, the German Federal Institute for Risk Assessment (BfRshort forGerman Federal Institute for Risk Assessment) conducted a health risk assessment of the results on the occurrence of mycotoxins in plant-based drinks from the study "Initial characterisation of selected plant-based drinks with regard to their quality and to their microbiological and chemical safety " conducted by the Max Rubner Institute (MRIshort forMax Rubner Institute) (MRIshort forMax Rubner Institute 2023). The results were contextualised with regard to their relevance for consumer exposureClose dialogexposureExposureTo glossary and an assessment was made as to whether the concentrations determined could pose a health risk to vulnerable groups. The BfRshort forGerman Federal Institute for Risk Assessment was unable to assess the extent to which the data collected by the MRIshort forMax Rubner Institute was representative of the German market and therefore recommended generating further data on the occurrence of mycotoxins in plant-based drinks.

The BfRshort forGerman Federal Institute for Risk Assessment collected the relevant data in 2024 and 2025 and, based on this data, prepared an updated risk assessment.

In its opinion 029/2024, the BfRshort forGerman Federal Institute for Risk Assessment conducted an assessment of the health risks posed by the intake of mycotoxins from the consumption of plant-based drinks for aflatoxin B1 (AFB1) in almond drinks, deoxynivalenol (DON) in oat drinks and T-2 and HT-2 toxins (T2/HT2) in oat drinks. The BfRshort forGerman Federal Institute for Risk Assessment collected further data in 2024 and 2025 and used this data to perform an updated risk assessment. A total of 162 plant-based drinks were tested for mycotoxins and plant toxins. The products tested were selected with the aim of obtaining a representative survey of the plant-based drinks available on the German market. Based on market research data, the products with the highest market shares were examined in order to achieve the highest possible market coverage for three types of plant-based drinks (oat drink, almond drink and soy drink). The plant-based drinks were analysed using a multi-analyte method, which, in addition to the mycotoxins already examined in the MRIshort forMax Rubner Institute study, enabled the determination of numerous other substances from the area of mycotoxins and plant toxins.

Based on the new data from the BfRshort forGerman Federal Institute for Risk Assessment, the health risk assessment of ochratoxin A (OTA) in soy drinks and almond drinks, for which no meaningful data was available in the MRIshort forMax Rubner Institute report, was supplemented for this opinion. The health risk assessment will thereby be limited to those substance-matrix combinations that have been detected most frequently and for which sufficient toxicological data are available as a basis for a proper risk assessment.

In addition, the health risk assessment is limited to children aged 0.5 to <6 years as a vulnerable consumer group. Compared to adults, and due to their lower body weight, children have a higher consumption and thus a higher exposure in relation to their body weight.

The most recent data on consumption for children is available from the KiESEL study (The Children’s Nutrition Survey to Record Food Consumption). As only a small number of children consumed plant-based drinks in the KiESEL study (n=38), cow's milk consumption is primarily used for the exposure assessment. This is based on the assumption that the corresponding plant-based drinks are used as an alternative to cow's milk and thus result in comparable consumption amounts.

A comparison with the actual consumption of plant-based drinks recorded in the KiESEL study suggests that using data on cow's milk consumption as a substitute for plant-based drink consumption could result in an overestimation. However, the differences in consumption levels are not significant. It should also be noted that sales of plant-based drinks are increasing, so it can be assumed that consumption of plant-based drinks has also increased since the field phase of the KiESEL study (2014-2017). However, the BfRshort forGerman Federal Institute for Risk Assessment does not have more recent consumption data to support this assumption.

The BfRshort forGerman Federal Institute for Risk Assessment has applied the margin of exposure (MOEshort forMargin of Exposure) concept to assess the health risks of AFB1 intake in almond drinks, as aflatoxins are genotoxic carcinogens for which no safe intake level can be derived. In its assessment of the health risks associated with the intake of AFB1 from the consumption of almond drinks based on realistic scenarios, the BfRshort forGerman Federal Institute for Risk Assessment concludes that in the case of long-term consumption of almond drinks containing AFB1 at concentrations determined by the BfRshort forGerman Federal Institute for Risk Assessment, the likelihood of health impairments in children aged 0.5 to <6 years is medium. If aflatoxins B2, G1 and G2 are also taken into account in the assessment of health risks, the BfRshort forGerman Federal Institute for Risk Assessment reaches the same conclusion.

Given the very low OTA concentrations determined by the BfRshort forGerman Federal Institute for Risk Assessment, realistic scenarios for the intake of OTA from the consumption of almond drinks or soy drinks by children aged 0.5 to <6 years result in margins of exposure (MOEshort forMargin of Exposure) that are considered to be of low concern for public health with regard to both the neoplastic and non-neoplastic effects of OTA.

In its assessment of the health risks associated with the intake of DON from the consumption of oat drinks based on realistic scenarios, the BfRshort forGerman Federal Institute for Risk Assessment concludes that in the case of both long-term and short-term consumption of oat drinks with DON at concentrations determined by the BfRshort forGerman Federal Institute for Risk Assessment, health impairments in children aged 0.5 to <6 years are unlikely. When the modified forms 3-acetyl-DON (3-Ac-DON), 15-Ac-DON and DON-3-glucoside (DON-3-Glu) are also taken into account in the assessment of health risks, the BfRshort forGerman Federal Institute for Risk Assessment reaches the same conclusion.

When conducting the health risk assessment of the intake of T2/HT2 alone through the consumption of oat drinks, the BfRshort forGerman Federal Institute for Risk Assessment concludes, based on realistic scenarios, that in the case of both short-term and long-term consumption of oat drinks containing T2/HT2 at concentrations determined by the BfRshort forGerman Federal Institute for Risk Assessment, the likelihood of health impairments in children aged 0.5 to <6 years is low.

However, it should be noted that, based on the average concentrations of T2/HT2 determined by the BfRshort forGerman Federal Institute for Risk Assessment, long-term high consumption of oat drinks by children aged 0.5 to <6 years already results in 82 % of the TDIshort forTolerable Daily Intake. It should be further noted that oat drinks are not the only source of T2/HT2 intake, but that T2/HT2 is also ingested through the consumption of other foods, in particular other oat products. In this opinion, the BfRshort forGerman Federal Institute for Risk Assessment assumes that plant-based drinks are used as an alternative to cow's milk. Accordingly, a scenario in which children aged 0.5 to < 6 years consume oat flakes together with oat drinks can be considered realistic. The consumption of oat drinks as an alternative to cow's milk would therefore represent an additional source of T2/HT2 intake, thereby increasing both the overall exposure to T2/HT2 and the likelihood of health impairments.

In addition, the BfRshort forGerman Federal Institute for Risk Assessment examined the plant-based drinks for plant toxins. In one soy drink product, the tropane alkaloids (TAs) atropine and scopolamine were repeatedly found at such high concentrations that the occurrence of health impairments must be considered. To this end, the BfRshort forGerman Federal Institute for Risk Assessment carried out individual case studies on the short-term intake of TAs from the consumption of the soy drink product with the highest TAs contamination and the soy drink product with the second-highest TAs contamination. Using the data for short-term consumption and the maximum TA content in soy drinks of 1,274 ng/kgshort forkilogram as determined by the BfRshort forGerman Federal Institute for Risk Assessment, this results in percentages of 100 % (average consumption) and 330 % (high consumption) of the ARfDshort forAcute Reference Dose. Exceeding the ARfDshort forAcute Reference Dose can generally lead to health impairments. In contrast, an analogous calculation using a TA concentration of 57 ng/kgshort forkilogram, which was the maximum determined by the BfRshort forGerman Federal Institute for Risk Assessment in the other soy drink samples, results in percentages of 4.5 % (average consumption) and 15 % (high consumption). Thus, the plant-based drinks examined (with the exception of the soy drink product mentioned above) do not indicate that the occurrence of TAs in plant-based drinks could pose a health risk to consumers.

The BfRshort forGerman Federal Institute for Risk Assessment also tested the plant-based drinks for lupanine (representative for a contamination with quinolizidine alkaloids – QAs). For classification purposes, the BfRshort forGerman Federal Institute for Risk Assessment carried out an exemplary worst-case scenario for the short-term intake of lupanine through high consumption (P95) of oat drinks. This resulted in a MOEshort forMargin of Exposure value of 16 when using the maximum lupanine concentration of 240 µgshort formicrogram/kgshort forkilogram determined by the BfRshort forGerman Federal Institute for Risk Assessment, and a correspondingly higher MOEshort forMargin of Exposure value of 1,300 when using a lupanine concentration at the 95th percentile (3 µgshort formicrogram/kgshort forkilogram). This means that even for these exemplary worst case scenarios, the MOEshort forMargin of Exposure values are well above 1 and thus within a range that, according to the EFSAshort forEuropean Food Safety Authority's assessment, does not pose any health concerns.

The BfRshort forGerman Federal Institute for Risk Assessment has generated a comprehensive data set on the occurrence of mycotoxins in the three best-selling plant-based drinks (oat, almond and soy). The results of this data collection will be used to update the BfRshort forGerman Federal Institute for Risk Assessment opinion 029/2024 ("Mycotoxins in plant-based drinks: more data required") of 25 June 2024.

For the update, the BfRshort forGerman Federal Institute for Risk Assessment will not conduct a health risk assessment for all substance-matrix combinations examined, but will limit itself to those cases that have been detected most frequently and for which sufficient toxicological data are available as a basis for a proper risk assessment. The risk assessment covers the following substance-matrix combinations:

In almond drinks

• Aflatoxin B1 (AFB1) and the sum of aflatoxins B1, B2, G1 and G2 (AFT)
• Ochratoxin A (OTA)

In soy drinks

• Ochratoxin A (OTA)

In oat drinks

• Deoxynivalenol (DON) and the sum of DON and its modified forms (3-Ac-DON, 15-Ac-DON and DON-3-Glu)
• The sum of toxins T-2 and HT-2 (T2/HT2)

In addition, the BfRshort forGerman Federal Institute for Risk Assessment also surveyed the concentrations of selected plant toxins (pyrrolizidine alkaloids, tropane alkaloids and lupanine, representative for quinolizidine alkaloids) in the plant-based drinks mentioned above. The classification of these concentrations can be found under point 3.6 of the opinion.

3.1.1 Hazard identification for AFB1 and the sum of aflatoxins B1, B2, G1 and G2

Aflatoxins belong to the group of mycotoxins and are mainly produced by the two fungal species Aspergillus flavus and Aspergillus parasiticus. Aflatoxins can be detected in various foods, such as tree nuts, peanuts, maize, spices and dried fruit (EFSAshort forEuropean Food Safety Authority 2020a).

Regulation (EU) 2023/915 sets maximum levels for aflatoxins (differentiated for aflatoxin B1 and the sum of aflatoxins B1, B2, G1 and G2) in various foods. Current maximum levels of 8.0 µgshort formicrogram/kgshort forkilogram (for AFB1) and of 10.0 µgshort formicrogram/kgshort forkilogram (for the sum of the aflatoxins B1, B2, G1 and G2) apply to almonds, placed on the market for the final consumer or for use as an ingredient in food, but there is no specific maximum level for aflatoxins in almond drinks. Regulation (EU) 2023/915 does, however, set a maximum level of 0.05 µgshort formicrogram/kgshort forkilogram for the main metabolite aflatoxin M1 in raw milk, heat-treated milk, and milk for the manufacture of milk-based products.

3.1.2 Hazard identification for OTA

OTA is the most prevalent mycotoxin within the ochratoxin group and has the highest toxic potential. OTA is mainly produced by fungi of the genera Aspergillus and Penicillium, in particular by the fungal species A. ochraceus, A. carbonarius and P. verrucosum. OTA can be detected in plant-based foods such as grains and grain products, spices, dried fruit and coffee, as well as in animal food products such as mature ham and hard cheese (EFSAshort forEuropean Food Safety Authority 2020b).

Regulation (EU) 2023/915 sets maximum levels for OTA in various foods. For example, a maximum level of 5 µgshort formicrogram/kgshort forkilogram applies to soybeans. There are no specific maximum levels for OTA in soy drinks and almond drinks according to Regulation (EU) 2023/915.

3.1.3 Hazard identification for DON and for the sum of DON and its modified forms

DON is one of the mycotoxins primarily produced by fungal species of the genus Fusarium. Human exposure occurs mainly through grains and grain products, in which DON can be detected almost ubiquitously (EFSAshort forEuropean Food Safety Authority 2017c).

Regulation (EU) 2023/915 sets maximum levels for DON in various foods. For example, a maximum level of 750 µgshort formicrogram/kgshort forkilogram applies to oats placed on the market for the final consumer. Regulation (EU) 2023/915 sets no specific maximum level for DON in oat drinks. Furthermore, modified forms of DON are not (yet) taken into account when setting maximum levels.

3.1.4 Hazard identification for T2/HT2

T-2 toxin and its main metabolite HT-2 toxin are mycotoxins that are primarily formed by Fusarium langsethiae, but also by some other Fusarium species. Human exposure occurs mainly through grains and grain products, with oats and oat products having the highest concentrations (EFSAshort forEuropean Food Safety Authority 2017b).

Regulation (EU) 2024/1038, which entered into force on 1 July 2024, established maximum levels for T2/HT2 in food for the first time. For example, a maximum level of 100 µgshort formicrogram/kgshort forkilogram applies to oats placed on the market for the final consumer and to oat flakes. There is no specific maximum level for T2/HT2 in oat drinks.

3.1.5 Plant-based drinks

Plant-based drinks are products that are advertised as plant-based milk alternatives and are further differentiated according to their ingredients. The BfRshort forGerman Federal Institute for Risk Assessment has examined almond drinks, oat drinks and soy drinks. Colloquially, plant-based drinks are also referred to as plant milk because they resemble milk in appearance and consistency and are commonly used by consumers as an alternative to milk. However, in the European Union, the term "milk" may only be used for products obtained by milking an udder (cow's, sheep's and goat's milk).

3.2.1 Hazard characterisation for AFB1 and the sum of aflatoxins B1, B2, G1 and G2

The International Agency for Research on Cancer (IARC) has classified aflatoxins in Group 1 (carcinogenic to humans) and last confirmed this classification in 2012 (IARC 2012). The carcinogenic potential of aflatoxins B1 and G1 was detected, while the results for the other two aflatoxins, B2 and G2, were less clear. The health risk assessment cites the potential to cause liver cancer as a critical effect, which is higher for aflatoxin B1 than for aflatoxin G1.

Aflatoxins have genotoxic and mutagenic properties. The European Food Safety Authority (EFSAshort forEuropean Food Safety Authority) therefore concluded that no dose without effect can be determined and that, accordingly, no health-based guidance values (HBGVs) can be established.

EFSAshort forEuropean Food Safety Authority therefore recommends using the margin of exposure (MOEshort forMargin of Exposure) concept for the health risk assessment of aflatoxins. The MOEshort forMargin of Exposure is the quotient of a suitable toxicological reference value and human exposure to the substance. For carcinogenic compounds, the benchmark dose lower confidence limit (BMDLshort forBenchmark Dose Lower Confidence Limit) is often used as a reference value. The BMDLshort forBenchmark Dose Lower Confidence Limit₁₀ is determined by modelling appropriate dose-effect relationship data and corresponds to the lower limit of the confidence interval of the dose that, in the case of a carcinogenic effect, is associated with an additional cancer risk of 10 % (benchmark dose response 10 %, BMR₁₀) compared to the control group. From a public health perspective, a margin of exposure (MOEshort forMargin of Exposure) of 10,000 and above is generally considered to be of low concern – but not harmless – for genotoxic carcinogens and is therefore given low priority for risk management measures.

EFSAshort forEuropean Food Safety Authority has identified a two-year carcinogenicity study in rats (Wogan et al.short foret alii (lat. "and others") 1974) with AFB1 as the critical study for the toxicity of aflatoxins. Based on the data on the induction of liver cell carcinomas in male rats, a benchmark dose lower confidence limit BMDLshort forBenchmark Dose Lower Confidence Limit₁₀ of 0.4 µgshort formicrogram/kgshort forkilogram bw per day was determined using benchmark modelling (EFSAshort forEuropean Food Safety Authority 2020a).

EFSAshort forEuropean Food Safety Authority also analysed the data on the potency of aflatoxins B1, B2, G1 and G2, but was unable to identify any clear quantitative differences. EFSAshort forEuropean Food Safety Authority therefore recommends that the same relative potency be assumed for aflatoxins B1, B2, G1 and G2 in a risk assessment (EFSAshort forEuropean Food Safety Authority 2020a).

The BfRshort forGerman Federal Institute for Risk Assessment has tested plant-based drinks for AFB1 as well as aflatoxins B2, G1 and G2, so that the health risk for both AFB1 alone and the sum of aflatoxins B1, B2, G1 and G2 is assessed below.

3.2.2 Hazard characterisation for OTA

Animal studies in rodents have shown that oral administration of OTA leads to hepatocellular and renal carcinomas. Furthermore, OTA has induced nephrotoxic and immunosuppressive effects in in vivo studies. It could not be concluded with certainty whether the effects observed in animal studies also lead to carcinogenicity in humans, so the IARC has classified OTA in Group 2B (possibly carcinogenic to humans) (IARC 1993).

OTA has a high binding affinity to plasma proteins, which leads to reduced excretion and, consequently, to comparatively long half-lives (5-6 days in pigs and up to 35 days in humans) and, in the case of chronic exposure, to an accumulation of OTA in humans and animals (EFSAshort forEuropean Food Safety Authority 2020b).

In 2020, EFSAshort forEuropean Food Safety Authority conducted a re-evaluation of the health risks associated with exposure to OTA through the consumption of food and concluded that there are significant uncertainties regarding the genotoxic mechanisms that lead to kidney tumours in animal studies. As it could not be conclusively clarified whether OTA induces direct or indirect genotoxic effects, EFSAshort forEuropean Food Safety Authority considered that the application of a threshold-based approach was no longer appropriate. As a result, the health-based guidance value (tolerable weekly intake, TWI) of 120 ng/kgshort forkilogram bw per week established by EFSAshort forEuropean Food Safety Authority in 2006 was suspended and the MOEshort forMargin of Exposure approach (see Section 3.2.1) was used instead for the assessment of the health risks.

For the neoplastic effects of OTA, EFSAshort forEuropean Food Safety Authority identified a two-year carcinogenicity study in rats (NTP 1989) as the critical study. Based on the data on the induction of kidney tumours in male rats, a benchmark dose lower confidence limit BMDLshort forBenchmark Dose Lower Confidence Limit₁₀ of 14.5 µgshort formicrogram/kgshort forkilogram bw per day was determined using benchmark modelling (EFSAshort forEuropean Food Safety Authority 2020b).

In addition, EFSAshort forEuropean Food Safety Authority has also established a toxicological reference value for the health risk assessment of the non-neoplastic effects of OTA. EFSAshort forEuropean Food Safety Authority has identified a 90-day study on sows as the most sensitive species as the critical study for this purpose (Krogh et al.short foret alii (lat. "and others") 1974). Based on the toxicologically most sensitive endpoint (early signs of renal toxicity such as microscopic changes in the renal tubules), a benchmark dose lower confidence limit BMDLshort forBenchmark Dose Lower Confidence Limit₁₀ of 4.73 µgshort formicrogram/kgshort forkilogram bw per day was determined (EFSAshort forEuropean Food Safety Authority 2020b).

With regard to public health, the EFSAshort forEuropean Food Safety Authority considers MOEshort forMargin of Exposure values of 10,000 and above for the neoplastic effects of OTA and MOEshort forMargin of Exposure values of 200 and above for the non-neoplastic effects of OTA to be of low concern  – but not harmless (EFSAshort forEuropean Food Safety Authority 2020b).

3.2.3 Hazard characterisation for DON and for the sum of DON and its modified forms

After oral intake of DON, both acute and chronic effects can occur, which can lead in particular to immunotoxic and developmental toxic effects. The mechanism of action involves binding to ribosomes, which leads to inhibition of protein biosynthesis. Pigs are particularly sensitive to DON and, depending on the level of oral exposure, show symptoms such as refusal to eat and vomiting, which is why DON is colloquially referred to as "vomitoxin".

The EFSAshort forEuropean Food Safety Authority has identified a two-year carcinogenicity study in mice (Iverson et al.short foret alii (lat. "and others") 1995) as the critical study for chronic DON intake. This study did not demonstrate a carcinogenic effect of DON, but did show reduced weight gain in the mice. Based on this effect as a critical endpoint, the EFSAshort forEuropean Food Safety Authority established a tolerable daily intake (TDIshort forTolerable Daily Intake) of 1 µgshort formicrogram/kgshort forkilogram bw per day (EFSAshort forEuropean Food Safety Authority 2017c).

In humans, primarily acute toxic effects have been described which manifest in nonspecific symptoms such as vomiting, diarrhoea, lower abdominal pain, headaches, and fever. As already described by Luo et al.short foret alii (lat. "and others") in 1987 in a study on an acute outbreak in China, these symptoms often occur just 30 minutes after consuming contaminated food. On the basis of this study, EFSAshort forEuropean Food Safety Authority has established an acute reference dose (ARfDshort forAcute Reference Dose) of 8 µgshort formicrogram/kgshort forkilogram bw per day (EFSAshort forEuropean Food Safety Authority 2017c).

The BfRshort forGerman Federal Institute for Risk Assessment points out that the two health-based guidance values do not refer to DON alone, but were established as group values for the sum of DON and its modified forms (3-acetyl-DON, 15-acetyl-DON, DON-3-glucoside) (EFSAshort forEuropean Food Safety Authority 2017c). This means that, in principle, the modified forms must also be taken into account in the health risk assessment.

The BfRshort forGerman Federal Institute for Risk Assessment also tested, in addition to DON, the plant-based drinks for 3-Ac-DON, 15-Ac-DON and DON-3-Glu, so that the health risk for both DON alone and the sum of DON and its modified forms is assessed below.

3.2.4 Hazard characterisation for T2/HT2

In in vivo studies, haematotoxic and myelotoxic effects and disrupted haematopoiesis were observed after administration of T-2 toxin. This is attributable to the T-2 toxin-mediated inhibition of protein biosynthesis (EFSAshort forEuropean Food Safety Authority 2017a). Because T-2 toxin is rapidly metabolised to HT-2 toxin, it is not possible to differentiate between the toxic effects of T-2 and HT-2 toxin and therefore the health-based guidance values were established for the sum of the toxins T-2 and HT-2.

In 2017, EFSAshort forEuropean Food Safety Authority reassessed the TDIshort forTolerable Daily Intake for T2/HT2, referring to a study by Rahman et al.short foret alii (lat. "and others") from 2014 as a critical study. In this subchronic 90-day study on rats, a reduction in total leukocyte count was observed and a correlation with haematotoxic effects from in vivo studies on other species was established. On the basis of the results of the study by Rahman et al.short foret alii (lat. "and others") (2014), a group TDIshort forTolerable Daily Intake of 0.02 µgshort formicrogram/kgshort forkilogram bw per day was established for the sum of T2/HT2 and their modified forms. The TDIshort forTolerable Daily Intake was established on the basis of a 10 % reduction in total leukocyte count, which is within the range of individual physiological variation and is not yet considered adverse (EFSAshort forEuropean Food Safety Authority 2017a).

In addition, EFSAshort forEuropean Food Safety Authority considered in its reassessment also the derivation of an ARfDshort forAcute Reference Dose for short-term exposure to T2/HT2. Acute in vivo studies on mink revealed that, on oral or intraperitoneal exposure, both T-2 and HT-2 toxin had emetic effects which were considered to be the most sensitive endpoint for acute exposure to T2/HT2. On the basis of the results of a study by Wu et al.short foret alii (lat. "and others") (2016), a group ARfDshort forAcute Reference Dose of 0.3 µgshort formicrogram/kgshort forkilogram bw per day was established for the sum of T2/HT2 and their modified forms. The ARfDshort forAcute Reference Dose is not established directly on the basis of emesis, but on a 10 % increase in the plasma level of the hormone 5-hydroxytryptamine (5 HT) and the peptide hormone PYY3-36, which are involved in the induction of vomiting (EFSAshort forEuropean Food Safety Authority 2017a).

The BfRshort forGerman Federal Institute for Risk Assessment points out that the two HBGVs are not limited to T2/HT2, but were established as group values for the sum of T2/HT2 and their modified forms (EFSAshort forEuropean Food Safety Authority 2017a). This means that, in principle, the modified forms must also be taken into account in the health risk assessment. Since the MRI’s investigations are limited to T2/HT2, the health risk below will also only be assessed for these two substances.

3.3.1 Data on the consumption of plant-based drinks and cow's milk as a substitute

3.3.1.1 Data basis for consumption

As an update to the VELS study, the BfRshort forGerman Federal Institute for Risk Assessment carried out a representative study throughout Germany, “KiESEL” (“The Children’s Nutrition Survey to Record Food Consumption”). The study was linked as a module to the Robert Koch Institute’s “German Health Interview and Examination Survey for Children and Adolescents” (“KiGGS Wave 2”).

A total of 1,104 children aged from six months up to and including five years participated in KiESEL over the period from 2014 to 2017. On the basis of an interview, the parents/guardians completed a questionnaire on general nutrition, nutrition in the first year of life and a Food Propensity Questionnaire on rarely consumed foods. Of these, 1,008 children or their parents also took part in the nutrition survey using a weighing/estimation record. The children’s food consumption was documented in a weighing record for three successive days and in a one-day weighing record on an independent day. In addition, out-of-home consumption (e.g. in a care setting) was acquired using a reduced estimation record (Nowak et al.short foret alii (lat. "and others") 2022).

In order to determine long-term consumption, all participants who had consumed cow’s milk (or plant-based drinks) on at least one day of the study had their consumption of the corresponding food group summed for the individual days of consumption and then the meanClose dialogmeanMeanTo glossary of all the days of the study was calculated. When determining short-term consumption, the maximum over all days of consumption was calculated instead. Children who were still partially breastfed were excluded from the evaluation. Consumption is presented in various age and gender groups. In addition, confidence intervals were determined non-parametrically using a bootstrap method.

3.3.1.2 Data on long-term consumption of cow's milk

The KiESEL study is the most recent representative consumption study for children in Germany. With a recording period up to 2017, it may not fully reflect current consumption trends, such as the consumption of plant-based drinks. For example, the KiESEL study documents only a small number of children who consumed plant-based drinks (n=38). The exposure assessment is therefore based primarily on the consumption of cow's milk. This is based on the assumption that the corresponding plant-based drinks are used as an alternative to cow's milk and therefore result in comparable consumption amounts. This assumption is particularly relevant for the increasing number of consumers who follow a vegan diet. The consumption of plant-based drinks and cow’s milk is compared in Section 3.3.1.4 to assess the assumption.

The results of long-term cow's milk consumption are presented in Table 1 . Children who did not consume cow's milk on at least one of the observation days were excluded from the evaluation. For this reason, the age group of 0.5 to under 1 year olds consists of only a small number of children who consume milk (n=15). In contrast, the proportion of children who had consumed cow's milk on at least one of the days was 85 % (n=811) across all age groups. Younger children consumed slightly more than older children, while there were no significant differences between boys and girls.

Across all age groups, children consumed a medianClose dialogmedianMedianTo glossary of 7.1 grams (g) per kilogram (kgshort forkilogram) of body weight (bw) per day (d). The high level of consumption, represented by the 95th percentile, amounts to 25.3 g/(kgshort forkilogram bw*d).

3.3.1.3 Data on short-term consumption of cow's milk

The short-term consumption of cow's milk by children aged 0.5 to <6 years is shown in Table 2. Across all age groups, the median was 12.8 g/(kgshort forkilogram bw*d) and the 95th percentile was 41.5 g/(kgshort forkilogram bw*d). Here, too, no significant differences between milk consumption in boys and girls could be demonstrated, while a reduction in consumption amounts with increasing age was observed when looking at the medians.

3.3.1.4 Comparison of data on cow's milk consumption with data on plant-based drink consumption

In order to be able to estimateClose dialogestimateEstimateTo glossary the extent to which the use of data on cow’s milk reflects the consumption of plant-based drinks, a comparison was made with children’s consumption data available from the KiESEL study. The median long-term consumption of plant-based drinks across all age groups (n=38) was 5.5 g/(kgshort forkilogram bw*d) (95 % CI 3.6–7.6 g/(kgshort forkilogram bw*d)) and thus below the consumption of cow’s milk of 7.1 g/(kgshort forkilogram bw*d) (95 % CI 6.3–7.8 g/(kgshort forkilogram bw*d)). However, since the confidence intervals for the consumption of cow’s milk and of plant-based drinks overlap, this difference was not significant. The same applies when considering the 95th percentile of long-term consumption, which was 13.2 g/(kgshort forkilogram bw*d) (95 % CI 11.7¬30.0 g/(kgshort forkilogram bw*d)) for the consumption of plant-based drinks compared to 25.3 g/(kgshort forkilogram bw*d) (95 % CI 23.2-29.9 g/(kgshort forkilogram bw*d)) for the consumption of cow’s milk.

The median short-term consumption of plant-based drinks was 11.9 g/(kgshort forkilogram bw*d) (95 % CI 8.7–15.2 g/(kgshort forkilogram bw*d)) and the 95th percentile was 20.6 g/(kgshort forkilogram bw*d) (95 % CI 18.2¬44.8 g/(kgshort forkilogram bw*d)). Comparison with the consumption of cow’s milk, for which median consumption amounts to 12.8 g/(kgshort forkilogram bw*d) (95 % CI 12.0-13.6 g/(kgshort forkilogram bw*d)) and the 95th percentile to 41.5 g/(kgshort forkilogram bw*d) (95 % CI 35.1-45.2 g/(kgshort forkilogram bw*d)), reveals differences that are likewise not significant.

The comparison of consumption data indicates that using cow’s milk consumption data as substitute for the consumption of plant-based drinks might be an overestimation. However, the differences between the consumption amounts are not significant. It should furthermore be taken into account that sales of plant-based drinks are rising and it can therefore be assumed that the consumption of plant-based drinks has also increased since the field phase of the KiESEL study (2014-2017). However, more recent consumption data to support this assumption are not available.

3.3.2 Data on the occurrence of mycotoxins in plant-based drinks

3.3.2.1 Sample planning

The aim of the sample planning was to obtain as representative a picture as possible of plant-based drinks on the German market. To this end, market data was purchased from the household panel of the company YouGov (formerly GfK). This panel represents a representative sample of consumers in Germany. The results contain the ten most important brands and products for different groups of plant-based drinks (e.g. oat drinks or almond drinks) together with their respective market share in litres.

The market data shows that the market for plant-based drinks is very heterogeneous. For oat drinks, the ten products with the highest market share cover only about 37 % of the total market. The products at the bottom of the ranking each have market shares of only 2 to 3 %, which suggests a very high number of different products. In the case of almond drinks, the 10 products with the highest market share cover around 70 % of the market, and in the case of soy drinks, the figure is 69 %. The products at the bottom of the rankings have market shares of approximately 3 % (almond drinks) and 4 % (soy drinks).

The high heterogeneity of the market was also confirmed by parallel research conducted in Mintel's Global New Products Database (GNPD) (MINTEL 2025). This database continuously incorporates new products entering the market. For the years 2020 to 2025 alone, 229 new products were listed here in the categories of oat drinks, soy drinks and almond drinks.

Based on the results, it was decided to aim for the highest possible market coverage, taking into account the available laboratory capacity. To this end, the ten products with the highest market share were selected for drinks made of oat, almond and soy. Furthermore, additional oat drinks from the three manufacturers with the largest market share were selected, as these combined have a market share of around 65 % according to market data, in order to increase market coverage as much as possible. Overall, however, it is unclear how large the total market coverage is.

In order to obtain an overview of the variability in concentrations, at least three batches from 2024 should be purchased from each selected product and, in a second phase at the beginning of 2025, two further batches from each. Furthermore, a list of plant-based drinks already tested at the MRIshort forMax Rubner Institute in 2022 was consulted for sampling (MRIshort forMax Rubner Institute 2023). Additional batches of each of the products listed there were purchased, provided the products were still available.

3.3.2.2 Sample purchase

Commercially available plant-based drinks based on oats, soy and almonds were purchased on the market and examined. The plant-based drinks were purchased in two periods: from October to November 2024 and from March to April 2025. The plant-based drinks were purchased at various retail outlets in Berlin, with the brand and variety of the product being taken into account in a manner representative of market data.

The purchasing team received lists containing information such as product name, brand, minimum purchase quantity and transport temperature. For each product, three packages from the same batch and a total minimum quantity of 1,000 mlshort formillilitre were purchased and assigned an internal code for tracking purposes. After receipt and registration at the BfRshort forGerman Federal Institute for Risk Assessment, the samples were stored unrefrigerated until they were handed over to the testing laboratory. For quality assurance and documentation of all information on the packaging, each product was photographed upon receipt and the product names, corresponding batch numbers and "best before" dates were checked. A total of 162 plant-based drinks were purchased, of which 92 in 2024 and 70 in 2025 were obtained. Table 3 shows the number of samples per year and type.

3.3.2.3 Further data sources

The BfRshort forGerman Federal Institute for Risk Assessment does not have any occurrence data on mycotoxin levels in plant-based drinks from the monitoring programmes of the German federal states ("Laender"). Plant-based drinks were examined in 2021 as part of a monitoring project, but only for the occurrence of elements.

3.3.2.4 Analysis of samples by the Federal Institute for Risk Assessment

A multi-analyte method for testing plant-based drinks was developed and fully validated at the National Reference Laboratory for Mycotoxins and Plant Toxins. Compared to previous methods, sensitivity has been significantly improved to achieve LOQs of well below 1 µgshort formicrogram/kgshort forkilogram in most cases.

In addition to the mycotoxins already investigated in the MRIshort forMax Rubner Institute study, the substances analysed (analytes) include numerous other substances from the area of mycotoxins and plant toxins. Validation was carried out in accordance with the requirements of Regulation (EU) 2023/2782 on sampling and analysis methods for the control of mycotoxin levels in food.

Overall, the method was validated for 92 mycotoxins and plant toxins. For two analytes (phomopsin A and nivalenol), validation for soy drinks was not possible, thus 90 analytes were validated in this matrix. A detailed overview of the substances examined with the validated detection and limits of quantification (LOQshort forLimit of quantification) can be found in Table 19 in the appendix.

After analysis of all 162 samples, a data set with 14,836 individual results was compiled. The assessment of health risks will be limited to those substance-matrix combinations that were detected most frequently and for which a toxicological assessment basis is available. In addition to the substance-matrix combinations subject to health risk assessment shown below, numerous other mycotoxins were detected in the plant-based drinks examined, for example enniatins in oat drinks or sterigmatocystin in all three types of plant-based drinks examined. A statistical evaluation of the entire data set can be found in the tables in Appendix 6.4 .

3.3.2.5 Comparison with the occurrence data from the Max Rubner Institute

Table 4 shows a comparison of the occurrence data collected by the BfRshort forGerman Federal Institute for Risk Assessment with the corresponding occurrence data from the MRIshort forMax Rubner Institute study (MRIshort forMax Rubner Institute 2023). Only those substance-matrix combinations that were used for the following exposure assessments are listed. Since the 95th percentile of the MRIshort forMax Rubner Institute data was not reported, it cannot be used for comparison. Furthermore, the MRIshort forMax Rubner Institute did not analyse aflatoxins B2, G1 and G2 or the modified forms of DON (3-acetyl-DON, 15-acetyl-DON, DON-3-glucoside), meaning that no comparative values from the MRIshort forMax Rubner Institute study are available for these total levels either.

Also, the MRIshort forMax Rubner Institute was evaluated using the modified lower bound approach, while the BfRshort forGerman Federal Institute for Risk Assessment applied both the modified lower bound approach and the upper bound approach. The difference between the two approaches lies in how values below the limit of detection and the limit of quantification (LOQshort forLimit of quantification) are handled. In the modified lower bound approach, values below the LODshort forLimit of detection are replaced with zero and values between the LODshort forLimit of detection and the limit of quantification (LOQshort forLimit of quantification) are replaced with the value for the LODshort forLimit of detection. In terms of statistical evaluation, the modified lower bound approach represents the best case, i.e. the reported statistical parameters represent the lower limit of the concentrations that are at least present in the collected data set. In contrast, in the upper bound approach, values below the limit of detection (LODshort forLimit of detection) are replaced with the LODshort forLimit of detection value and values between the limit of detection and the limit of quantification (LOQshort forLimit of quantification) are replaced with the LOQshort forLimit of quantification value. In terms of statistical evaluation, the upper bound approach thus represents the worst case, i.e. the reported statistical parameters represent the upper limit of the maximum concentrations present in the collected data set.

The data from the modified lower bound approach is used to compare the occurrence data collected by the BfRshort forGerman Federal Institute for Risk Assessment with that of the MRIshort forMax Rubner Institute. The MRIshort forMax Rubner Institute has stated its values in ng/l, the BfRshort forGerman Federal Institute for Risk Assessment in ng/kgshort forkilogram. Since the density of commercially available plant-based drinks is very close to that of water, the units are equated for the purposes of this opinion. As can be seen from Table 4 , the concentrations determined by the BfRshort forGerman Federal Institute for Risk Assessment for AFB1 in almond drinks, at 8 ng/kgshort forkilogram (MW) and 55 ng/kgshort forkilogram (maximum concentration) respectively, are slightly lower than those determined by the MRIshort forMax Rubner Institute (MW = 18.1 ng/l and maximum value = 130.0 ng/l). In contrast, the data collected by the BfRshort forGerman Federal Institute for Risk Assessment for DON in oat drinks, with a mean value of 2,900 ng/kgshort forkilogram and a maximum value of 35,000 ng/kgshort forkilogram, show significantly higher concentrations than those determined by the MRIshort forMax Rubner Institute (MW = 691.8 ng/l and maximum value = 5,457.5 ng/l). For T2/HT2 in oat drinks, the data collected by the BfRshort forGerman Federal Institute for Risk Assessment and the MRIshort forMax Rubner Institute are at a comparable level, showing mean values of 470 ng/kgshort forkilogram (BfRshort forGerman Federal Institute for Risk Assessment) and 397.2 ng/l (MRIshort forMax Rubner Institute) and maximum values of 2,300 ng/kgshort forkilogram (BfRshort forGerman Federal Institute for Risk Assessment) and 2,146.5 ng/l (MRIshort forMax Rubner Institute).

Furthermore, with respect to oat drinks, it was investigated whether the concentrations in the samples taken in 2024 differed from those in the samples taken in 2025 against the background of the newly introduced maximum levels for the sum of T-2 and HT-2 toxins. The mean value for 2024 was 600 ng/kgshort forkilogram in the modified lower bound and 740 in the upper bound. For the samples from 2025, the values are 360 ng/kgshort forkilogram (modified lower bound) and 570 ng/kgshort forkilogram (upper bound). To decide whether these lower values for 2025 are significant, the Peto-Peto test (Peto&Peto, 1927) was used on the basis of the censored values. This revealed a significant difference between the two years. However, the values for both years are below the values collected by the MRIshort forMax Rubner Institute for samples from 2023 (see Table 4 ), which suggests that the introduction of maximum levels is not the cause of this difference. Regular testing of oats has shown that mycotoxin concentrations are subject to annual fluctuations, mostly due to weather conditions. For this reason, the values from both years have been summarised below.

3.3.3 Exposure assessment

3.3.3.1 Methodological approach

Long-term exposure was determined by multiplying the data for long-term consumption (median/50th percentile and 95th percentile) by the mean and 95th percentile of the concentrations, respectively. The use of the mean of the concentrations represents the exposure of children who are exposed to random concentrations over a longer period of time. In addition, the use of the 95th percentile of the concentrations represents a worst-case calculation for long-term intake. In this scenario, it is assumed that only plant-based drinks with mycotoxin concentrations in the range of the 95th percentile are consumed over a long period of time. In the previous opinion, the maximum concentration was used, deviating from the usual procedure of the BfRshort forGerman Federal Institute for Risk Assessment, as the 95th percentile of the concentrations was not reported in the MRIshort forMax Rubner Institute report. Accordingly, only the 95th percentile of the concentrations is listed for the BfRshort forGerman Federal Institute for Risk Assessment data in Table 4.

To determine short-term exposure, the data for short-term consumption (median/50th percentile and 95th percentile) were multiplied exclusively by the 95th percentile of the concentrations, as in this case the assumption of short-term consumption of a plant-based drink with a high mycotoxin concentration represents a realistic scenario.

As can be seen from Table 4 , the statistical evaluation shows only minor differences between the upper bound approach and the modified lower bound approach. This is due to the low proportion of left-censored data (concentrations below the detection and LOQshort forLimit of quantification limits) in the substance-matrix combinations considered. Due to these minor differences between the upper bound approach and the modified lower bound approach, only the occurrence data from the upper bound approach was used for the exposure assessments and the subsequent risk characterisations in accordance with the precautionary principle. For comparison, the exposure assessments using the occurrence data from the modified lower bound approach can be found in the tables in the appendix.

As already stated at the beginning of this opinion, the exposure assessment below will be limited to the substance-matrix combinations AFB1 and OTA in almond drinks, OTA in soy drinks, and DON and T2/HT2 in oat drinks.

3.3.3.2 Estimation of exposure to AFB1 and the sum of aflatoxins B1, B2, G1 and G2 from the consumption of almond drinks

Table 5 shows long-term exposure to AFB1 and the sum of aflatoxins B1, B2, G1 and G2 from the consumption of almond drinks. At average consumption and mean AFB1 concentrations, the exposure of children aged 0.5 to <6 years is 0.09 ng/(kgshort forkilogram bw^[1] d). If, on the other hand, high consumption is assumed, the intake increases to 0.33 ng/(kg bwd). Assuming the 95th percentile of AFB1 concentrations, the intake is 0.25 ng/(kgshort forkilogram bw*d) for average consumption and 0.90 ng/(kgshort forkilogram BW*d) for high consumption. If aflatoxins B2, G1 and G2 are also taken into account in the exposure assessment, the long-term exposure to the sum of aflatoxins increases to 0.15 ng/(kgshort forkilogram bw*d) (average consumption) or 0.52 ng/(kgshort forkilogram bw*d) (high consumption) and, when using the 95th percentile of the concentrations, to 0.39 ng/(kgshort forkilogram bw*d) (average consumption) or 1.40 ng/(kgshort forkilogram bw*d) (high consumption).

3.3.3.3 Estimated exposure to OTA from the consumption of almond and soy drinks

Table 6 shows the long-term exposure to OTA from the consumption of almond drinks and soy drinks. For almond drinks, the exposure of children aged 0.5 to <6 years at average consumption and mean levels is 0.22 ng/(kgshort forkilogram bw*d). If, on the other hand, high consumption is assumed, the intake increases to 0.77 ng/(kgshort forkilogram bw*d). Assuming the 95th percentile of concentrations, the intake is 0.31 ng/(kgshort forkilogram bw*d) for average consumption and 1.10 ng/(kgshort forkilogram bw*d) for high consumption.

In comparison, the exposure of children aged 0.5 to <6 years with average consumption of soy drinks with mean concentrations is 0.40 ng/(kgshort forkilogram bw*d). If, on the other hand, high consumption is assumed, the intake increases to 1.40 ng/(kgshort forkilogram bw*d). If the 95th percentile of concentrations is assumed, the intake is 1.30 ng/(kgshort forkilogram bw*d) for average consumption and 4.60 ng/(kgshort forkilogram bw*d) for high consumption.

3.3.3.4 Estimated exposure to DON and the sum of DON and its modified forms from the consumption of oat drinks

The results for short- and long-term exposure to DON and the sum of DON and its modified forms (3-Ac-DON, 15-Ac-DON, DON-3-Glu) from the consumption of oat drinks are presented in Table 7. The long-term exposure of children aged 0.5 to <6 years is 21 ng/(kgshort forkilogram bw*d) ( average consumption) or 76 ng/(kgshort forkilogram bw*d) (high concentration) at mean DON concentrations. In the case of the 95th percentile of DON concentrations, long-term exposure is 65 ng/(kgshort forkilogram bw*d) (average consumption) or 231 ng/(kgshort forkilogram bw*d) (high consumption). If the modified forms 3-Ac-DON, 15-Ac-DON, DON-3-Glu are also taken into account in the exposure assessment, the long-term exposure to the sum of DON and its modified forms increases to 42 ng/(kgshort forkilogram bw*d) (average consumption) or 150 ng/(kgshort forkilogram bw*d) (high consumption) when using the mean concentrations, and to 95 ng/(kgshort forkilogram bw*d) (average consumption) or 340 ng/(kgshort forkilogram bw*d) (high consumption) when using the 95th percentile of the concentrations.

When considering short-term consumption, the exposure of children aged 0.5 to <6 years to DON is 120 ng/(kgshort forkilogram bw*d) (average consumption) or 380 ng/(kgshort forkilogram bw*d) (high consumption). When modified forms of DON are also taken into account, short-term exposure increases to 170 ng/(kgshort forkilogram bw*d) (average consumption) and 550 ng/(kgshort forkilogram bw*d) (high consumption).

3.3.3.5 Estimated exposure to T2/HT2 from the consumption of oat drinks

Table 8 shows the results of long-term and short-term exposure assessments for the intake of T2/HT2 from the consumption of oat drinks. The long-term exposure of children aged 0.5 to <6 years is 4.6 ng/(kgshort forkilogram bw*d) (average consumption) or 16 ng/(kgshort forkilogram bw*d) (high consumption), with an assumption of mean concentrations. Using the 95th percentile of concentrations, the intake levels for long-term exposure are 11 ng/(kgshort forkilogram bw*d) (average consumption) and 37 ng/(kgshort forkilogram bw*d) (high consumption).

The percentages of health-based guidance values and the calculation of the margin of exposure (MOEshort forMargin of Exposure) refer to the exposure assessments in Chapter 3.3 and are summarised in Table 9 to Table 13 .

3.4.1 Risk characterisation for AFB1 and for the sum of aflatoxins B1, B2, G1 and G2 in almond drinks

As already explained in Chapter 3.2.1 , the margin of exposure (MOEshort forMargin of Exposure) concept must be applied to assess the health risks posed by the intake of AFB1 and the sum of aflatoxins B1, B2, G1 and G2, as aflatoxins are genotoxic carcinogens for which no safe intake level can be established. For this purpose, a benchmark dose lower confidence limit (BMDLshort forBenchmark Dose Lower Confidence Limit) of 400 ng/(kgshort forkilogram bw*d) is used as the toxicological reference value, which was established from a two-year carcinogenicity study in rats.

The BfRshort forGerman Federal Institute for Risk Assessment emphasises that this toxicological reference value does not represent a health-based guidance value (HBGVshort forHealth-Based Guidance Value), but merely serves to prioritise risk management measures. From a public health perspective, a margin of exposure (MOEshort forMargin of Exposure) of 10,000 and above is generally considered to be of low concern – but not harmless – for genotoxic carcinogens and is therefore given low priority for risk management measures.

Using the data on long-term exposure to AFB1 according to Table 5 in Chapter 3.3.3.2 , MOEshort forMargin of Exposure values of between 450 and 4,400 are obtained, depending on the scenario under consideration, which are thus well below an MOEshort forMargin of Exposure value of 10,000 for all the scenarios under consideration (Table 9 ). If aflatoxins B2, G1 and G2 are also taken into account in the exposure assessment, the MOEshort forMargin of Exposure values are even lower, ranging between 290 and 2,800 (Table 9 ).

In its assessment of the health risks associated with the intake of AFB1 from the consumption of almond drinks based on realistic scenarios, the BfRshort forGerman Federal Institute for Risk Assessment concludes that long-term consumption of almond drinks containing AFB1 at concentrations determined by the BfRshort forGerman Federal Institute for Risk Assessment may result in a medium likelihood of health impairments in children aged 0.5 to <6 years. When aflatoxins B2, G1 and G2 are also taken into account in the assessment of health risks, the BfRshort forGerman Federal Institute for Risk Assessment reaches the same conclusion.

When conducting an assessment of the health risks posed by chronic intake of AFB1 and the sum of aflatoxins B1, B2, G1 and G2, it must also be taken into account that almond drinks are not the only source of aflatoxins, but that aflatoxins are also ingested when consuming other foods. In this assessment, it is assumed that almond drinks are consumed as an alternative to cow's milk. When aflatoxins are transferred from feed to milk, they are metabolised, which results in cow's milk mainly containing the metabolite aflatoxin M1 (AFM1), which has about ten times less toxic potential than AFB1. In contrast, aflatoxins that may be present in almonds used for the production of almond drinks are not metabolised to AFM1 during the manufacturing process. The BfRshort forGerman Federal Institute for Risk Assessment was able to detect AFB1 in the vast majority of the almond drinks examined (31 out of 39 samples). The average concentrations detected were 13 ng/kgshort forkilogram, with a maximum concentration of 55 ng/kgshort forkilogram. For AFM1 in raw milk, heat-treated milk and milk for the manufacture of milk-based products, a maximum concentration of 50 ng/kgshort forkilogram is set according to Regulation (EU) 2023/915. With the assumption that the concentrations of AFB1 in plant-based drinks and AFM1 in cow's milk are comparable, consuming plant-based drinks as an alternative to cow's milk would increase the proportion of AFB1 in the overall exposure and at the same time reduce the proportion of AFM1. As AFB1 has a higher toxic potential than AFM1, this would increase the likelihood of health impairments.

The present risk characterisations based on representative data thus support the conclusions of the previous opinion that the intake of AFB1 and the sum of aflatoxins B1, B2, G1 and G2 from the consumption of almond drinks may pose a health risk to vulnerable consumer groups.

3.4.2 Risk characterisation for OTA in almond drinks and soy drinks

As already described in Section 3.2.2, EFSAshort forEuropean Food Safety Authority recommends applying the margin of exposure (MOEshort forMargin of Exposure) concept as a precautionary measure for the assessment of health risks from OTA intake, as it could not be conclusively clarified whether OTA induces direct or indirect genotoxic effects. A BMDLshort forBenchmark Dose Lower Confidence Limit₁₀ of 14.5 µgshort formicrogram/(kgshort forkilogram bw*d) is used as the toxicological reference value for the neoplastic effects of OTA, which was established from a two-year carcinogenicity study of rats (EFSAshort forEuropean Food Safety Authority 2020b).

The BfRshort forGerman Federal Institute for Risk Assessment once again emphasises that this toxicological reference value does not represent a health-based guidance value (HBGVshort forHealth-Based Guidance Value), but merely serves to prioritise risk management measures. From a public health perspective, a margin of exposure (MOEshort forMargin of Exposure) of 10,000 and above is generally considered to be of low concern – but not harmless – for genotoxic carcinogens and is therefore given low priority for risk management measures.

Using the data on long-term exposure to OTA according to Table 6 in Chapter 3.3.3.3 results for the consumption of almond drinks in margin of exposure (MOEshort forMargin of Exposure) values in the range of 13,000 to 67,000 – depending on the scenario considered –and, for intake from the consumption of soy drinks, MOEshort forMargin of Exposure values in the range of 3,100 to 37,000 (Table 10). This means that, with one exception (worst-case scenario for long-term high intake of soy drinks with exclusively high OTA concentrations), the MOEshort forMargin of Exposure values for all scenarios considered are above an MOEshort forMargin of Exposure value of 10,000 and thus within a range that EFSAshort forEuropean Food Safety Authority considers to be of low concern with regard to the neoplastic effects of OTA.

However, as the mechanism of the genotoxic effect of OTA could not be conclusively clarified, EFSAshort forEuropean Food Safety Authority recommends additionally conducting a health risk assessment of the non-neoplastic effects of OTA and also applying the margin of exposure (MOEshort forMargin of Exposure) concept for this purpose. A BMDLshort forBenchmark Dose Lower Confidence Limit₁₀ of 4.73 µgshort formicrogram/(kgshort forkilogram bw*d) was established from a 90-day study in sows as a toxicological reference value for this purpose. For the risk characterisation of the non-neoplastic effects of long-term exposure to OTA, EFSAshort forEuropean Food Safety Authority considers MOEshort forMargin of Exposure values of 200 and above to be of low concern for public health (EFSAshort forEuropean Food Safety Authority 2020b).

Using the data on long-term exposure to OTA according to Table 6 in Chapter 3.3.3.3, MoEshort forMargin of Exposure values in the range of 4,400 to 22,000 are obtained for the non-neoplastic effects of OTA from the consumption of almond drinks, depending on the scenario considered, and MOEshort forMargin of Exposure values in the range of 1,000 to 12,000 for intake from the consumption of soy drinks (Table 11). Thus the MOEshort forMargin of Exposure values for all scenarios considered are well above an MOEshort forMargin of Exposure value of 200 and thus within a range that EFSAshort forEuropean Food Safety Authority considers to be of low concern with regard to the non-neoplastic effects of OTA.

The BfRshort forGerman Federal Institute for Risk Assessment detected OTA in the majority of almond drinks (33 out of 39 samples) and soy drinks (23 out of 29 samples). These high detection rates (especially when compared to the MRIshort forMax Rubner Institute studies) are mainly due to the high sensitivity of the method used by the BfRshort forGerman Federal Institute for Risk Assessment. The mean concentrations for almond drinks were 31 ng/kgshort forkilogram with a maximum concentration of 110 ng/kgshort forkilogram, and for soy drinks they were 55 ng/kgshort forkilogram with a maximum concentration of 290 ng/kgshort forkilogram. Given the very low OTA concentrations determined by the BfRshort forGerman Federal Institute for Risk Assessment, realistic scenarios for the intake of OTA from the consumption of almond drinks or soy drinks by children aged 0.5 to <6 years result in margins of exposure that are considered to be of low concern for public health in terms of both the neoplastic and non-neoplastic effects of OTA.

When assessing the health risks posed by chronic OTA intake, it should also be taken into account that OTA is one of the mycotoxins that can enter the food chain via a variety of sources. Plant-based drinks thus represent an additional source of OTA intake from food consumption that has not been taken into account in exposure assessment to date and may contribute to overall exposure to OTA.

3.4.3 Risk characterisation for DON and for the sum of DON and its modified forms in oat drinks

For the assessment of the health risks from the intake of DON and for the sum of DON and its modified forms from the consumption of oat drinks, the TDIshort forTolerable Daily Intake of 1,000 ng/(kgshort forkilogram bw*d) was used as the health-based guideline value for long-term exposure and the ARfDshort forAcute Reference Dose of 8,000 ng/(kgshort forkilogram bw*d) was used for short-term exposure.

Using the data on long-term exposure to DON according to Table 7 in Chapter 3.3.3.4 results in percentages in the range of 2.1 % to 23.1 % of the TDIshort forTolerable Daily Intake depending on the scenario considered (Table 12). If the modified forms 3-Ac-DON, 15-Ac-DON and DON-3-Glu are also taken into account in the exposure assessment, the percentages of the TDIshort forTolerable Daily Intake increase to between 4.2 % and 34 % (Table 12).

Using a similar approach results in percentages for short-term DON intake of 1.5 % (average consumption) and 4.7 % (high consumption) of the ARfDshort forAcute Reference Dose (Table 12). When the modified forms 3-Ac-DON, 15-Ac-DON and DON-3-Glu are also taken into account in the exposure assessment, the percentages increase to 2.1 % (average consumption) and 6.9 % (high consumption) (Table 12).

In its assessment of the health risks associated with the intake of DON from the consumption of oat drinks based on realistic scenarios, the BfRshort forGerman Federal Institute for Risk Assessment concludes that both long-term and short-term consumption of oat drinks containing DON at the concentrations determined by the BfRshort forGerman Federal Institute for Risk Assessment are unlikely to cause health impairments in children aged 0.5 to <6 years. The BfRshort forGerman Federal Institute for Risk Assessment reaches the same conclusion when the modified forms 3-Ac-DON, 15-Ac-DON and DON-3-Glu are also taken into account in the assessment of health risks.

The present risk characterisations based on representative data thus support the conclusions of the previous opinion.

3.4.4 Risk characterisation for T2/HT2 in oat drinks

For the assessment of health risks from the intake of T2/HT2 from the consumption of oat drinks, the TDIshort forTolerable Daily Intake of 20 ng/(kgshort forkilogram bw*d) was used as the health-based guidance value for long-term exposure and the ARfDshort forAcute Reference Dose of 300 ng/(kgshort forkilogram bw*d) for short-term exposure.

Using the data on long-term exposure to T2/HT2 according to Table 8 in Chapter 3.3.3.5 results in percentages in the range of 23 % to 190 % of the TDIshort forTolerable Daily Intake depending on the scenario considered (Table 13).

Using a similar approach for short-term intake results in percentages of 6 % (average consumption) and 20 % (high consumption) of the ARfDshort forAcute Reference Dose (Table 13).

In its assessment of the health risks associated with the short-term intake of T2/HT2 from the consumption of oat drinks based on realistic scenarios, the BfRshort forGerman Federal Institute for Risk Assessment concludes that short-term consumption of oat drinks containing T2/HT2 concentrations determined by the BfRshort forGerman Federal Institute for Risk Assessment results in a low likelihood of health impairments in children aged 0.5 to <6 years.

However, when conducting an assessment of the health risks associated with the intake of T2/HT2, it must also be taken into account that oat drinks are not the only source of T2/HT2, but that T2/HT2 can also be ingested from the consumption of other oat products, such as oat flakes. In a scenario where all oat products consumed have T2/HT2 concentrations equal to the maximum level of 100 µgshort formicrogram/kgshort forkilogram applicable since 1 July 2024, the ARfDshort forAcute Reference Dose for children aged 1 to <6 years could be exceeded in the case of short-term consumption (154–371 % of the ARfDshort forAcute Reference Dose). A possible additional intake of T2/HT2 from the consumption of oat drinks was not taken into account in this scenario.

In its assessment of the health risks associated with long-term intake of T2/HT2 from the consumption of oat drinks based on realistic scenarios, the BfRshort forGerman Federal Institute for Risk Assessment concludes that in the case of long-term average consumption of oat drinks with T2/HT2 concentrations equal to the mean concentrations determined by the BfRshort forGerman Federal Institute for Risk Assessment, the likelihood of health impairments in children aged 0.5 to <6 years is low.

In contrast, long-term high consumption solely of oat drinks with T2/HT2 concentrations equal to the mean concentrations determined by the BfRshort forGerman Federal Institute for Risk Assessment by children aged 0.5 to <6 years already results in 82 % of the TDIshort forTolerable Daily Intake. It should also be noted that oat drinks are not the only source of T2/HT2 intake, but that T2/HT2 is also ingested from the consumption of other foods, in particular other oat products. In this opinion, the BfRshort forGerman Federal Institute for Risk Assessment assumes that plant-based drinks are used as an alternative to cow's milk. Accordingly, a scenario in which children aged 0.5 to < 6 years consume oat flakes together with oat drinks can be considered realistic. The consumption of oat drinks as an alternative to cow's milk would therefore represent an additional source of T2/HT2 intake, thereby increasing both the overall exposure to T2/HT2 and the likelihood of health impairments.

The BfRshort forGerman Federal Institute for Risk Assessment was able to detect T2/HT2 in all oat drinks tested (86 out of 86 samples). The average concentration was 650 ng/kgshort forkilogram, with a maximum concentration of 2,300 ng/kgshort forkilogram. The present risk characterisations based on representative data thus support the conclusions of the previous opinion that, for children aged 0.5 to <6 years as a vulnerable consumer group, the consumption of oat drinks as an alternative to cow's milk may represent an additional source of intake of T2/HT2.

With a survey period from 2014 to 2017, the KiESEL study is the most up-to-date representative consumption study for children in this age group in Germany. However, changes in consumption behaviour since the survey period cannot be ruled out. This applies in particular to the consumption quantities for plant-based drinks (not used for the exposure assessment), as the market has changed significantly since the survey period. However, it is not possible on the basis of this data to conclusively assess the assumption that the consumption of plant-based drinks can be adequately described by cow’s milk consumption data. The BfRshort forGerman Federal Institute for Risk Assessment assumes that the uncertainties regarding the consumption of plant-based drinks have only a minor impact on the result of the exposure assessment, as it is plausible to assume that the consumption behaviour for cow's milk and plant-based drinks is similar. However, the calculations performed here are based on the indirect assumption that the entire consumption of cow's milk is replaced by only one type of plant-based drink. For consumers who divide their consumption between different plant-based drinks, the scenarios presented here overestimate the intake from individual plant-based drinks.

Compared to adults, children, due to their lower body weight, have a higher consumption and thus a higher exposure in relation to their body weight. A self-determined vegetarian/vegan diet typically only develops in adolescence and puberty, so the proportion of consumers and perhaps also the quantity of plant-based drinks consumeWith a survey period from 2014 to 2017, the KiESEL study is the most up-to-date representative consumption study for children in this age group in Germany. However, changes in consumption behaviour since the survey period cannot be ruled out. This applies in particular to the consumption quantities for plant-based drinks (not used for the exposure assessment), as the market has changed significantly since the survey period. However, it is not possible on the basis of this data to conclusively assess the assumption that the consumption of plant-based drinks can be adequately described by cow’s milk consumption data. The BfRshort forGerman Federal Institute for Risk Assessment assumes that the uncertainties regarding the consumption of plant-based drinks have only a minor impact on the result of the exposure assessment, as it is plausible to assume that the consumption behaviour for cow's milk and plant-based drinks is similar. However, the calculations performed here are based on the indirect assumption that the entire consumption of cow's milk is replaced by only one type of plant-based drink. For consumers who divide their consumption between different plant-based drinks, the scenarios presented here overestimate the intake from individual plant-based drinks.

Compared to adults, children, due to their lower body weight, have a higher consumption and thus a higher exposure in relation to their body weight. A self-determined vegetarian/vegan diet typically only develops in adolescence and puberty, so the proportion of consumers and perhaps also the quantity of plant-based drinks consumed could be higher in this age group than in younger children. However, since the calculations were based on cow's milk as a substitute rather than the actual consumption of plant-based drinks, the intake calculations can be considered sufficiently conservative for all age groups.

It should also be noted that the exposure assessments and the associated risk characterisations are limited to mycotoxin intake from the consumption of the respective plant-based drinks, meaning that the total exposure through food consumption (as already explained in Chapters 3.4.1 and 3.4.4 ) may be significantly higher.

The occurrence data used for the risk assessment are subject to measurement uncertainty. A relative expanded measurement uncertainty of 50 % was estimated for the results collected by the BfRshort forGerman Federal Institute for Risk Assessment (expansion factor k=2).d could be higher in this age group than in younger children. However, since the calculations were based on cow's milk as a substitute rather than the actual consumption of plant-based drinks, the intake calculations can be considered sufficiently conservative for all age groups.

It should also be noted that the exposure assessments and the associated risk characterisations are limited to mycotoxin intake from the consumption of the respective plant-based drinks, meaning that the total exposure through food consumption (as already explained in Chapters 3.4.1 and 3.4.4 ) may be significantly higher.

The occurrence data used for the risk assessment are subject to measurement uncertainty. A relative expanded measurement uncertainty of 50 % was estimated for the results collected by the BfRshort forGerman Federal Institute for Risk Assessment (expansion factor k=2).

As described under 3.3.2.4 , the plant-based drinks were analysed using an LC-MS/MS-based multi-method that covered selected plant toxins in addition to mycotoxins. The results on the occurrence of these plant toxins are classified below.

3.6.1 Tropane alkaloids

Tropane alkaloids (TA) are secondary plant compounds with a high occurrence in certain plants such as henbane, thorn apple and belladonna. More than 200 different TA have been identified to date (BfRshort forGerman Federal Institute for Risk Assessment 2013). Data on the occurrence of TA in food and feed and on the toxicity of TA are only available to a limited extent and primarily for atropine and scopolamine, which are also used medicinally. These compounds are known to affect the heart rate and central nervous system even at low doses (BfRshort forGerman Federal Institute for Risk Assessment 2013). Atropine is a racemate consisting of equal parts of the two enantiomers (-)-hyoscyamine and (+)-hyoscyamine, whereby only (-)-hyoscyamine is toxicologically relevant. The same ultimately applies to the (-)-enantiomer of scopolamine. However, since only the (-)-enantiomers are selectively formed in the plant anyway (EFSAshort forEuropean Food Safety Authority 2013), differentiation is not necessary when testing plant-based foods, as it can be assumed that "atropine" and "scopolamine" are exclusively the toxicologically relevant (-)-enantiomers.

When conducting an assessment of the health risks of TAs, the focus is on acute toxic effects. Based on a human study of healthy young adults, in whom higher doses resulted in the occurrence of a reduction in heart rate and effects on the central nervous system, such as dizziness, headaches and nausea, EFSAshort forEuropean Food Safety Authority has derived a NOAEL of 0.16 µgshort formicrogram/kgshort forkilogram bw per day. Applying an additional safety factor of 10 for interindividual variability in the population, this results in an acute reference dose (ARfDshort forAcute Reference Dose) of 0.016 μg/kgshort forkilogram bw per day. The ARfDshort forAcute Reference Dose applies to the sum of (-)-hyoscyamine and (-)-scopolamine (BfRshort forGerman Federal Institute for Risk Assessment 2013).

The BfRshort forGerman Federal Institute for Risk Assessment tested the plant-based drinks for TAs. One soy drink product stood out as having a higher TA concentration than the other products. This product was sampled a total of four times at different points during the investigation and, with TA concentrations of 271, 705, 856 and 1,274 ng/kgshort forkilogram, had the four highest TA concentrations of all the plant-based drinks tested. In comparison, the TA concentrations in the other 25 soy drink samples examined ranged from below the limit of detection to a maximum of 57 ng/kgshort forkilogram.

In order to classify these findings, the BfRshort forGerman Federal Institute for Risk Assessment conducted individual case studies on the short-term intake of TAs from the consumption of the soy drink product with the highest TAs contamination and the soy drink product with the second-highest TAs contamination. Using the data for short-term consumption according to Table 2 and the maximum TA content in soy drinks determined by the BfRshort forGerman Federal Institute for Risk Assessment of 1,274 ng/kgshort forkilogram, this results in percentages of 100 % (average consumption) and 330 % (high consumption) of the ARfDshort forAcute Reference Dose. Exceeding the ARfDshort forAcute Reference Dose can, in principle, lead to the occurrence of health impairments. In contrast, an analogous calculation using a TA concentration of 57 ng/kgshort forkilogram, which was the maximum determined by the BfRshort forGerman Federal Institute for Risk Assessment in the other soy drink samples, results in percentages of 4.5 % (average consumption) and 15 % (high consumption). Thus, the plant-based drinks examined (with the exception of the one reported soy drink product) do not indicate that the occurrence of TAs in plant-based drinks could pose a health risk to consumers.

3.6.2 Chinolizidine alkaloids

Chinolizidine alkaloids (QAs) are bitter-tasting secondary plant compounds that occur naturally in lupin seeds (Khan et al.short foret alii (lat. "and others") 2015; Wink 2019). In total, more than 170 structurally different QAs are known, with lupanine, hydroxylupanine, albin, sparteine, lupinine and angustifoline being the main representatives in food production, depending on the variety used (BfRshort forGerman Federal Institute for Risk Assessment 2025).

When conducting an assessment of the health risks associated with QAs, the focus is on acute toxic effects such as pupil dilation, dizziness, nausea, dry mouth, stomach pain, vomiting, diarrhoea or heart problems. Based on human data on sparteine, EFSAshort forEuropean Food Safety Authority has derived a toxicological reference value of 0.16 mgshort formilligram/kgshort forkilogram bw per day for considerations relating to the margin of exposure (MOEshort forMargin of Exposure). This corresponds to the lowest effective dose of sparteine in pharmaceutical applications. For the other QAs, a comparable effect and potency to sparteine is assumed, so that a group assessment is carried out for QAs under the assumption of dose additivity. If a comparison with the estimated short-term intake of QAs results in MOEshort forMargin of Exposure values of 1 and above, EFSAshort forEuropean Food Safety Authority considers that there are no health concerns (EFSAshort forEuropean Food Safety Authority 2019).

The BfRshort forGerman Federal Institute for Risk Assessment has tested plant-based drinks for lupanine (representative for contamination with QAs). Lupanine was only quantified in oat drinks, and in that case in 16 of 86 samples (19 %). For classification purposes, the BfRshort forGerman Federal Institute for Risk Assessment conducted an exemplary worst-case analysis for the short-term intake of lupanine from high consumption (P95) of oat drinks. This resulted in a MOEshort forMargin of Exposure value of 16 when using the maximum lupanine concentration of 240 µgshort formicrogram/kgshort forkilogram determined by the BfRshort forGerman Federal Institute for Risk Assessment, and a correspondingly higher MOEshort forMargin of Exposure value of 1,300 when using a lupanine concentration equal to the 95th percentile of the concentrations (3 µgshort formicrogram/kgshort forkilogram). This means that even for these exemplary "worst case" scenarios, the MOEshort forMargin of Exposure values are well above 1 and thus within a range that, according to EFSAshort forEuropean Food Safety Authority, does not pose any health concerns.

3.6.3 Pyrrolizidine alkaloids

Pyrrolizidine alkaloids (PAs) are a large group of compounds produced mainly by plants, but also by fungi and bacteria (Robertson & Stevens 2017). Several hundred PAs and their N-oxides are known to date (Wiedenfeld et al.short foret alii (lat. "and others") 2008). The primary target organ for toxic effects in humans is the liver. The genotoxic-carcinogenic effects of 1,2-unsaturated PAs are considered the most sensitive endpoint (EFSAshort forEuropean Food Safety Authority 2017d). Regulation (EU) 2023/915 sets maximum levels for the sum of 21 PAs and 14 other co-eluting PAs in various foods.

The BfRshort forGerman Federal Institute for Risk Assessment tested plant-based drinks for the legally regulated PAs, but was unable to quantify them in any sample (all samples <LOQshort forLimit of quantification). Based on the concentrations determined by the BfRshort forGerman Federal Institute for Risk Assessment, the occurrence of PAs in plant-based drinks can therefore be considered of little concern to public health.

For some substance-matrix combinations that were also frequently detected but for which there is currently insufficient toxicological data, it was not possible to make a meaningful assessment of the health risks. This applies, for example, to enniatins in oat drinks and sterigmatocystin in all three types of plant-based drinks examined.

Toxicological studies specifically for enniatins are currently being conducted as part of the PARC project (European Partnership for the Assessment of Risks from Chemicals), which was explicitly designed to close data gaps in risk assessment. In parallel, EFSAshort forEuropean Food Safety Authority has already been mandated to conduct an assessment of the risks to animal and human health from the occurrence of enniatins in feed and food (M-2024-00047). The deadline for the EFSAshort forEuropean Food Safety Authority opinion is 30 September 2026.

BfRshort forGerman Federal Institute for Risk Assessment (2013) opinion No. 035/2014 "High tropane alkaloid levels in cereal products: Health impairments are possible in individuals with heart problems" dated 13 November 2013

BfRshort forGerman Federal Institute for Risk Assessment (2024) opinion 029/2024 "Mycotoxins in plant-based drinks: more data required" dated 25 June 2024

BfRshort forGerman Federal Institute for Risk Assessment (2025) opinion 035/2025 "Alkaloids and allergies: Current data on health risks from lupin seeds in food" dated 17 September 2025

EFSAshort forEuropean Food Safety Authority (European Food Safety Authority) (2006) Opinion of the Scientific Panel on Contaminants in the Food Chain on a request from the Commission related to ochratoxin A in food. EFSAshort forEuropean Food Safety Authority Journal 2006; 4(6):365; 56 pp.

EFSAshort forEuropean Food Safety Authority (European Food Safety Authority) (2013) Scientific Opinion on Tropane alkaloids in food and feed. EFSAshort forEuropean Food Safety Authority Journal 2013; 11(10):3386; 113 pp.

EFSAshort forEuropean Food Safety Authority (European Food Safety Authority) (2017a) Appropriateness to set a group health-based guidance value for T-2 and HT-2 toxin and its modified forms. EFSAshort forEuropean Food Safety Authority Journal 2017; 15(1):4655; 53 pp.

EFSAshort forEuropean Food Safety Authority (European Food Safety Authority) (2017b) Human and animal dietary exposure to T-2 and HT-2 toxins. EFSAshort forEuropean Food Safety Authority Journal 2017; 15(8):4972; 57 pp.

EFSAshort forEuropean Food Safety Authority (European Food Safety Authority) (2017c) Risks to human and animal health related to the presence of DON and its acetylated and modified forms in food and feed. EFSAshort forEuropean Food Safety Authority Journal 2017; 15(9):4718; 345 pp.

EFSAshort forEuropean Food Safety Authority (European Food Safety Authority) (2017d) Risks for human health related to the presence of pyrrolizidine alkaloids in honey, tea, herbal infusions and food supplements. EFSAshort forEuropean Food Safety Authority Journal 2017; 15(7):4908; 34 pp.

EFSAshort forEuropean Food Safety Authority (European Food Safety Authority) (2019) Scientific opinion on the risks for animal and human health related to the presence of quinolizidine alkaloids in feed and food, in particular in lupins and lupin-derived products. EFSAshort forEuropean Food Safety Authority Journal 2019; 17(11):5860; 113 pp.

EFSAshort forEuropean Food Safety Authority (European Food Safety Authority) (2020a) Risk assessment of aflatoxins in food. EFSAshort forEuropean Food Safety Authority Journal 2020; 18(3):6040; 112 pp.

EFSAshort forEuropean Food Safety Authority (European Food Safety Authority) (2020b) Risk assessment of ochratoxin A in food. EFSAshort forEuropean Food Safety Authority Journal 2020; 18(5):6113; 150 pp.

IARC (International Agency for Research on Cancer) (1993) Ochratoxin A. IARC monographs on the evaluation of carcinogenic risks to humans. Some naturally occurring substances: food items and constituents, heterocyclic aromatic amines and mycotoxins, 56, 489-521.

IARC (International Agency for Research on Cancer) (2012) Aflatoxins. IARC monographs on the evaluation of carcinogenic risks to humans. Chemical Agents and Related Occupations. A review of Human Carcinogens, 100F, 225-248.

Iverson F, Armstrong C, Nera E, Truelove J, Fernie S, Scott P, Stapley R, Hayward S and Gunner S (1995) Chronic feeding study of deoxynivalenol in B6C3F1 male and female mice. Teratogenesis, Carcinogenesis and Mutagenesis, 15, 283-306.

Khan MK, Karnpanit W, Nasar-Abbas SM, Huma Z, Jayasena V (2015) Phytochemical composition and bioactivities of lupin: a review. International Journal of Food Science and Technology, 50, 2004-2012.

Krems C, Bauch A, Götz A, Heuer T, Hild A, Möseneder J, Brombach C (2006) Methoden der Nationalen Verzehrsstudie II.Ernährungs Umschau, 53(2), 44-50.

Krogh P, Axelsen NH, Elling F, Gyrd-Hansen N, Hald B, Hyldgaard-Jensen J, Larsen AE, Madsen A, Mortensen HP, Moller T, Petersen OK, Ravnskov U, Rostgaard M, Aalund O (1974) Experimental porcine nephropathy. Changes of renal function and structure induced by ochratoxin A-contaminated feed. Acta Pathologica, Microbiologica, et Immunologica Scandinavica [A], Supplement 246, 1-21.

Luo XY, Lishort forlithium YW, Wen SF and Hu X (1987) Food poisoning caused by scabby wheat and the detection of Fusarium mycotoxins. Journal of Hygiene Research, 16, 33-37.

MINTEL (2025), Mintel Global New Products Database, accessed on 13 October 2025 via External Link:https://www.mintel.com/de/produkte/gnpd/

MRIshort forMax Rubner Institute (Max Rubner Institute) (2023) Initial characterisation of selected plant-based drinks in terms of their quality and microbiological and chemical safety. Final report (in German) (External Link:https://www.mri.bund.de/de/pflanzendrink-bericht)

Nowak N, Diouf F, Golsong N, Höpfner T, Lindtner O (2022) KiESEL – The Children’s Nutrition Survey to Record Food Consumption for the youngest in Germany. BMC Nutrition, 8 (1), 64.External Link: https://doi.org/10.1186/s40795-022-00527-6

NTP (National Toxicology Programme) (1989) Toxicology and Carcinogenesis Studies of Ochratoxin A (CAS No. 303-47-9) in F344/N Rats (Gavage Studies). National Toxicology Programme, Technical Reports Series, 358, 1-142.

Peto R, Peto J (1972) Asymptotically Efficient Rank Invariant Test Procedures., Journal of the Royal Statistical Society Series A (General), 135(2).

Rahman S, Sharma AK, Singh ND, Telang AG, Azmi S, Prawez S (2014) Clinico-haematological changes in T-2 toxicosis in Wistar rats. Indian Journal of Veterinary Pathology, 38:22-28.

Robertson J, Stevens K (2017) Pyrrolizidine alkaloids: occurrence, biology, and chemical synthesis. Natural Product Reports, 34:62-89.

Commission Regulation (EU) 2024/1038 of 9 April 2024 amending Regulation (EU) 2023/915 as regards maximum levels for T-2 and HT-2 toxins in food. OJ L 1038, 10.4.2024, p. 1.

Commission Regulation (EU) 2023/915 of 25 April 2023 on maximum levels for certain contaminants in food and repealing Regulation (EC) No 1881/2006. OJ L 119, 5.5.2023, p. 103.

Wiedenfeld H, Roeder E, Bouraul T, Edgar JA (2008) Pyrrolizidine alkaloids. Structure and toxicity, V & R Unipress, Göttingen.

Wink M (2019) Quinolizidine and Pyrrolizidine Alkaloid Chemical Ecology - a Mini-Review on Their Similarities and Differences. Journal of Chemical Ecology, 45:109-115.

Wogan GN, Paglialunga S, Newberne PM (1974) Carcinogenic effects of low dietary levels of aflatoxin B1 in rats. Food and Cosmetics Toxicology, 12:681-685.

Wu W, Zhou H, Bursian SJ, Link JE, Pestka JJ (2016) Emetic responses to T-2 toxin, HT-2 toxin and emetine correspond to plasma elevations of peptide YY3-36 and 5-hydroxytryptamine. Archives of Toxicology, 90:997-1007.