Chemical Contaminants in Plant Based Protein Survey

9.1. A survey of plant protein products was carried out in two phases. Samples were purchased in Scotland or on-line at two time points. Twenty-six samples were purchased each time, these consisted of products such as plant protein powders, vegan sausages and burgers, shakes and protein bars made from different plant proteins. Samples in Phase 1 were analysed for regulated mycotoxins, tropane alkaloids and acrylamide, a subset were analysed for ergot alkaloids. Phase 2 samples were analysed for regulated mycotoxins, a broader range of Fusarium mycotoxins, ergot alkaloids, citrinin, Alternaria toxins, plant alkaloids (tropanes and pyrrolizidine), and a subset were analysed for erucic acid.   9.2. All analyses were carried out using accredited or in-house validated methods that met required quality control parameters. In Phase 1 one sample of soy powder contained a total aflatoxin level of 4.94 µg/kg. This sample also contained OTA (4.15 µg/kg). Another sample of pea powder contained 9.33 µg/kg OTA. While there are no MLs for protein products specifically, the levels of OTA and aflatoxins are at or above MLs in place for comparable foods. 

9.3. Alternaria toxins were found in over half the samples tested. There are no regulations in force for Alternaria toxins, however, these findings are similar to results reported in other published studies (6-10). Residues of other Fusarium toxins were found at low levels, except fusaric acid, which was found in 16 out of 26 samples. The highest levels reported were 118.4 µg/kg and >1000 µg/kg although the latter was an indicative result. This mycotoxin is difficult to analyse for; the analytical recovery was low which added to the uncertainty about the result. There are no MLs for fusaric acid and little information about its toxicological relevance. Therefore, it is not possible to make comparisons or comment on the significance of these 2 higher values. 

9.4. Overall, the results are reassuring, for the majority of samples tested, mycotoxins and plant toxins were not detected, despite the fact that low LOQs were achieved. For the mycotoxins found, the concentrations measured were below comparable MLs, except for two samples. The study provides valuable results to address a data gap about the occurrence of mycotoxins and plant toxins in vegetarian and plant based foods purchased in Scotland and can be used for risk and exposure assessments for plant based foods.  

Thanks to Food Standards Scotland for purchasing the samples. Thanks also to the following staff at Fera who contributed to this project: Jacob Harrison, Adam Tramasuer, Stephen Chapman, Verity Caddie and Joanna Stratton.