Investigation of Fusarium mycotoxins in UK barley and oat production

Executive Summary

This four-year project started in 2002 to ascertain the effects of agronomic practices on the concentration of fusarium mycotoxins in UK barley and oats over a number of seasons.  One hundred samples both of barley and oats were collected each year at harvest, together with agronomic details, and analysed for ten trichothecenes (including DON, nivalenol, HT2 and T2), zearalenone and moniliformin. 

The project anticipated the introduction by the European Commission (EC) of legislative limits for the fusarium mycotoxins, deoxynivalenol (DON) and zearalenone in cereals and cereal products intended for human consumption in July 2006.  A combined limit for HT2 toxin and T2 toxin (HT2+T2) will be introduced in the near future.

Fusarium mycotoxins are produced as a result of the disease fusarium ear blight (panicle blight in oats) caused by Fusarium species.  The most important ear blight pathogens on cereals worldwide are F. graminearum and F. culmorum which produce DON and zearalenone.  The vast majority of ear blight research and surveys of mycotoxin occurrence have been conducted on wheat as this is the most economically important small grain cereal worldwide and it is the most susceptible cereal to ear blight infection. 

The incidence and concentration of most fusarium mycotoxins, including DON and zearalenone, were low in both barley and oats compared to values for wheat.  This indicates that with current agronomic practices and varieties, wheat is the most susceptible host to F. culmorum and F. graminearum with barley and oats having considerably lower levels.  Concentrations of DON and zearalenone were below legislative limits for both barley and oats over the four year period 2002-2005.

The incidence and concentration of HT2 and T2 in UK barley samples was similar to UK wheat with ca. 1% of samples exceeding a combined concentration of 100 ppb.  The highest concentration was 138 ppb HT2+T2, which may, or may not exceed the legal limit if set at 100 ppb depending on the measurement of uncertainty with the assay used.

Regression analysis failed to identify relationships between fusarium mycotoxin concentrations in barley.  This is probably due to the low number of positive samples and the low concentration of these mycotoxins in positive samples.  Modelling of HT2+T2 concentration against the agronomy of barley failed to identify an effect of any agronomic factor other than year and region.

The incidence and concentration of HT2 and T2 were high in UK oats with quantifiable concentrations in 92% of samples and a combined concentration (HT2+T2) of 570 ppb for all samples analysed from 2002 to 2005.  The concentration of HT2+T2 was modelled against agronomic practices applied to each field.  Year, region, practice (organic or conventional), previous crop, cultivation and variety all had statistically significant effects on HT2+T2 concentration in oats. 

There was a degree of multicolinearity (ie related trends between different agronomic factors) within the observational data in that many conventional farmers grew the variety Gerald after another cereal, usually wheat; whereas organic farmers were more likely to grow other varieties after a non-cereal.  Consequently it could be identified that organic samples had a significantly lower HT2+T2 content compared to conventional samples and that this was partly due to organic growers growing Gerald and not following a cereal as frequently as conventional growers. 

Analysis indicated that one or more factors not included in the model, which differed between organic and conventional practice, also had an impact on HT2+T2 concentrations.  One possible difference is rotation, with organic growers tending to use longer, less cereal intense rotations. 

Statistical tests of the predictive quality of the model indicated it may not be a good predictor of new observations.  This indicates that the model should be used to formulate hypotheses as to the role of agronomic factors which can be quantified in field experiments under controlled conditions, rather than to predict the mycotoxin content in commercial samples based on known agronomy.

There was a significant interaction between year and region, which is probably due to fluctuation in weather between years and regions.  There was no trend from North to South, as seen for DON in wheat, which would indicate that the temperature difference across the UK does not limit HT2 and T2 production in oats.  Oat samples with more than 500 ppb HT2+T2 were detected in all regions of the UK at similar frequencies.

Highest concentrations of HT2+T2 occurred in oat crops grown after a previous cereal crop.  Oat samples from fields following a non-cereal and ploughing had significantly lower HT2+T2 than oat crops after wheat, barley or oats.  Oat samples from fields following a non-cereal and min-till had a HT2+T2 concentration equivalent to oat crops after a cereal.  This suggests that crop debris is important in the epidemiology of HT2+T2 producing Fusarium species. 

There were significant differences in the HT2+T2 content of different UK varieties.  Of the five varieties with sufficient samples to include within the analysis, Gerald, the most popular oat variety in recent years, had the highest HT2+T2 content.  Analysis of the HT2+T2 content of oat samples from the HGCA Recommended List trials allowed all current UK varieties to be compared from replicated field trials.  Results agreed with the observational data, with Gerald having the highest content of current varieties. 

HT2+T2 content of spring oat variety trial samples were consistently lower than winter oat samples and there was no significant difference between spring oat varieties tested in 2005.  Naked oat varieties tended to have a lower HT2+T2 content compared to conventional (covered) oat varieties.  Naked oats have a loose hull which is removed during harvesting.  Analysis of HT2+T2 content of two agronomy trials with different seed  and nitrogen rates indicated that they had no significant effect on HT2+T2 content of harvested oats. 

High levels of HT2 and T2 were detected in UK oats.  The combined HT2+T2 median, mean and maximum were 213, 570 and 9990 ppb respectively.  The previous European Commission limit for discussion was 500 ppb HT2+T2.  Thirty percent of samples in this study would have exceeded this limit; in each year of this study, between 18 to 50 percent of conventional oat samples would have exceeded this limit.  Prior to this study there was very limited data as to the concentration of HT2+T2 in oats worldwide.  In recent years (2002-2006) high HT2+T2 levels have occurred in northern European countries.

There was a good correlation between concentrations of all the type A trichothecenes detected (HT2, T2, T2 triol and neosolaniol).   These mycotoxins are likely to be produced by the same Fusarium species within the same metabolic pathway, and can be considered as co-contaminants.  There appeared to be some mutual exclusion between HT2, DON and nivalenol indicating that these mycotoxins are produced by different Fusarium species, which either actively compete with one another or have different environmental requirements.

All oats used for human consumption are de-hulled; the resulting groats are further processed into oat products and the hulls are pelleted for inclusion in animal feed.  De-hulling experiments were conducted to identify the impact of processing on the mycotoxin content of oats.  High levels of reduction, greater than 90%, were identified in an initial experiment of four samples, with a corresponding high level of HT2+T2 in the hulls.  A second, larger experiment of 66 samples showed a wider range of reduction (58-98%, average 89%), however this may have been due to sampling error as smaller samples were de-hulled (100 g compared to 500 g in the first experiment). 

This experiment identified no significant effect in the reduction during de-hulling of variety, groat content or the initial mycotoxin content of the oat sample.  A recent experiment on the impact of industrial processing on HT2+T2 content of oats found consistently high reductions of more than 90% from oats to groats and corresponding high levels in the pelleted hulls.  The impact of de-hulling explains the difference in mycotoxin content of oats at harvest as identified in this project and the low concentrations of HT2+T2 detected in retail oat products as found in a recent survey conducted in 2003 by the FSA.