View Report Details

Collection of honey samples potentially contaminated with pyrrolizidine alkaloids (PAs) from ragwort and borage and determination of PAs therein along with the an assessment of the stability of these compounds during storage of honey

Project Code: T01037

30/03/2007

LGC
Gormley, L

Pyrrolozidine alkaloids (PAs) are natural toxins produced by many plants, including ragwort and borage. There is a risk that they could contaminate honey if bees forage on PA-containing flowers. There has never been a recorded case of PA poisoning from honey.

This project aimed to investigate this risk, by producing honey under conditions likely to encourage PA contamination from both borage and ragwort. Such conditions amounted to normal production methods for specialty “borage honey”, but in the case of ragwort were artificial conditions designed to model a worst-case scenario.

Quantitative chemical analysis of contaminants requires the use of analytical standards (purified material of known concentration). Such standards are not available for the majority of the PAs. Production of standards was outside of the scope of this project, which was intended as a preliminary study to identify if the risk from PAs in honey warranted a full quantitative investigation. This study was therefore designed so that, although PA concentrations in the honey could not be quantified, they could be compared:
i) relatively from one honey sample to another; and
ii) relative to the amount of PAs in a fixed weight of plant material.

Without analytical standards, it cannot be certain that the analytical method and instrument settings were suitable for detecting all of the PAs of interest.

Results – Ragwort

The “worst-case” conditions used for this study would be unlikely to be replicated in practice, as large populations of ragwort are now very rare in the UK. Even under these conditions, it was very difficult to produce ragwort-tainted honey. There was little evidence of ragwort pollen in any of the honey. There was no difference between the PA profile in honey from “ragwort” sites compared to any other sites. There were elevated levels of one PA (seneciphylline-N-oxide) in samples from one site when hives were examined individually, but conditions at this site were extremely different than those that would be used for commercial honey production. There were also elevated PA levels in samples of honey abstracted from another site early during the production season, but these had fallen again by the time the honey was harvested at the end of the season.

Results – Borage

Honey produced from “borage” sites contained a significant amount of borage pollen. Chemical analysis identified a PA that could be either intermidine or lycopsamine. This PA was elevated in the honey from “borage” sites. If some samples in which no PA was detected were disregarded, there was a strong correlation between this PA and the borage pollen content of each honey. If these samples were not disregarded, there was no correlation. The worst-case assumption in terms of consumer risk would be to disregard these samples and assume a correlation. Also for worst-case modeling, the extraction of this PA from the herb used as a reference was assumed to by 100% (100% yields from such extractions and purifications are rare). In this worst-case, 10 g of high quality “borage honey” would contain the same amount of this PA as approximately 1.4 g of dried borage herb.

Main File

Some of the files on this site may be in a format that your computer can't read. However, you can download Readers and Viewers for the following document types below: