EFSA assesses health impacts of pyrrolizidine alkaloids in food and feed
The European Food Safety Authority (EFSA) has published a scientific opinion on the presence in food and feed of pyrrolizidine alkaloids (PAs), toxins produced naturally by several plant species, most of which are weeds. Sources of food and feed may be contaminated with PA toxins through contact with these plants. Experts from EFSA’s Panel on Contaminants in the Food Chain (CONTAM) identified a number of PAs of importance as potential contaminants in food and feed and concluded that there is a possible health concern for some high consumers of honey, the only food for which data on PA levels were available. The opinion also reaffirms EFSA’s previous work on PAs in feed, stating that the likelihood of animals being at risk from these toxins is low.
The CONTAM Panel stated that one class of PAs, known as 1,2-unsaturated PAs, may act as genotoxic carcinogens in humans (that is they may cause cancer and damage DNA, the genetic material of cells). The Panel’s experts concluded therefore that it was not appropriate to establish a Tolerable Daily Intake (TDI) and decided instead to use the Margin of Exposure (MOE) approach† to assess the potential risk from exposure to this class of PAs in food.
Although there might be other sources of PA exposure, due to lack of data the CONTAM Panel was not able to quantify exposure from food other than honey. (PAs are found in honey as a result of foraging by bees). EFSA’s experts used data on honey consumption across Europe to calculate the MOE for different consumers. EFSA’s Scientific Committee has previously concluded based on animal studies that for genotoxic carcinogens a MOE of 10,000 or higher is of low concern for public health. This is the case for adults for whom the MOE for consumption of honey contaminated with 1,2-unsaturated PAs is generally above 10,000. However, for toddlers and other children who are high consumers of honey, the MOE can be as low as 1200 and 3900 respectively, indicating a possible health concern. Additionally, for individuals who regularly eat unblended honey from certain small-scale producers, exposure to PAs could be up to twice that of people who consume honey industrially produced for retail.
In its conclusions, the CONTAM Panel supports ongoing efforts to collect more data on the PAs identified in its opinion and potentially found in feed and relevant foods, including milk, eggs and meat as well as in herbal dietary supplements prepared from PA-containing plants. As current detection methods are limited, the Panel also recommends a larger and more diverse set of certified reference standards for PAs. More data on the geographical and botanical origin of PAs is also needed to understand where PA concentrations might be highest. Finally, there is a need for toxicological data relating to the PAs most commonly found in honey to better understand any potential health risks for consumers.
An estimated 6,000 plant species worldwide may contain pyrrolizidine alkaloids (PAs). PAs are mainly found in the distantly related angiosperm families of the Boraginaceae (‘Forget-me-nots’), Asteraceae (the Senecioneae and Eupatorieae tribes of the daisy family) and Fabaceae (the Crotalaria genus commonly known as rattlepods). The PA-content in food and feed depends on a large number of factors, including species and organ of PA-producing plant, harvest, storage and extraction procedures.
The Margin of Exposure (MOE) is a tool used by risk assessors to characterise the risk from exposure to carcinogenic and/or genotoxic substances in food or feed. The MOE is a ratio of two factors which assesses for a given population: the dose at which a small but measurable adverse effect is first observed and the level of exposure to the substance considered. The higher the MOE, the lower the potential health risk is for consumers.
Due to the absence of toxicological data for most 1,2-unsaturated PAs, available data on a type of liver cancer caused by lasiocarpine, one of the most toxic 1,2-unsaturated PAs reported, were used for the MOE estimation. The CONTAM Panel considered this a conservative approach that was likely to account for exposure to other PAs that may also be present in honey.