Distribution of Fusarium mycotoxins in UK wheat mill fractions.

The EU has set maximum limits for the Fusarium mycotoxins, deoxynivalenol (DON) and zearalenone (ZON). The maximum permitted level decreases from unprocessed wheat, through intermediary products, e.g. flour, to finished products such as bakery goods and breakfast cereals. It is, therefore, important to understand the effects of processing on the mycotoxin distribution in mill fractions. Between 2004 and 2007, samples were taken at commercial flour mills at various points in the milling process and analysed for trichothecenes and ZON. Samples with a range of mycotoxin concentrations harvested in 2004 and 2005 were processed in a pilot mill and the mycotoxins in the different mill fractions quantified. In the commercial samples, DON was the predominant mycotoxin with highest levels detected in the bran fraction. Analysis of the pilot mill fractions identified a significant difference between the two years and between mycotoxins. The proportion of DON and nivalenol in the mill fractions varied between years. DON and nivalenol were higher in flour fractions and lower in bran and offal in samples from 2004 compared to samples from 2005. This may be a consequence of high rainfall pre-harvest in 2004 resulting in movement of these mycotoxins within grains before harvest. There was no significant difference in the distribution of ZON within mill fractions between the two years. For DON, higher concentrations in the grain resulted in a greater proportion of DON within the flour fractions. Understanding the factors that impact on the fractionation of mycotoxins during milling will help cereal processors to manufacture products within legislative limits.

Occurrence of Fusarium mycotoxins in maize imported into the UK, 2004-2007.

This study examined a total of 82 consignments of French and Argentinean raw maize as received at maize mills in the UK between 2004 and 2007. Samples were analysed for deoxynivalenol (DON), nivalenol (NIV), other trichothecenes, zearalenone (ZON), and fumonisins B(1), B(2), and B(3) (FB(1), FB(2), and FB(3)) using fully validated analytical methods with limits of quantification of 10 microg kg(-1) for DON, NIV, and each fumonisin mycotoxin and 3 microg kg(-1) for ZON. All samples except two containing fumonisins met the European Commission statutory maximum permissible levels for DON, ZON, and FB(1) + FB(2) as operating in 2007. The maximum concentrations found for DON, NIV, ZON, and FB(1) + FB(2) were 444, 496, 165 and 5002 microg kg(-1), respectively. Fumonisins were detected in almost every sample with 65% of Argentinean maize containing more than 1000 microg kg(-1) of FB(1) + FB(2). In contrast, ZON was not detectable in almost 50% of consignments. During this period there was a distinct difference in mycotoxin concentrations between harvests and geographic origin. Flint maize from Argentina usually contained lower concentrations of DON and related trichothecenes and higher levels of fumonisins than maize from France, although concentrations of fumonisins up to 2000 microg kg(-1) or greater occurred in samples from both regions. The incidence and concentrations of fumonisins were similar to those in a similar previous survey, while zearalenone concentrations were lower. The distribution of mycotoxins in multi-hold ships was also investigated showing that fumonisins were much more evenly distributed than DON, thus indicating their general level in the ship as a whole. The effect of cleaning regimes was found to be very variable, especially for DON, ranging from no removal of mycotoxins to greater than 50% in some instances, but was not related to concentration. Evidence here suggests that while cleaning is essential for removing foreign bodies before milling, it cannot be used as a reliable tool for reducing mycotoxins.

Fusarium mycotoxins in milling streams from the commercial milling of maize imported to the UK, and relevance to current legislation.

A study in three large commercial UK maize mills showed that Fusarium mycotoxins, such as deoxynivalenol, zearalenone and fumonisins present at mill intake, are distributed in milling streams approximately according to their occurrence in the maize seed structure. Fractions derived from the endosperm tended to contain the lowest levels of mycotoxins. Concentrations of mycotoxins within the endosperm are also related to the particle size. However, the products derived from the embryo or outer seed layers contained the highest mycotoxin levels being concentrated up to five times or more, although these components are normally used for animal feed or industrial use. The general pattern of mycotoxin distribution found when milling French and Argentinean maize was similar, although very variable, and it is concluded that this variability stems from different milling strategies used at each mill and from the nature and condition of each consignment of maize. Mycotoxins in maize grits (particle sizes >500 microm) were usually reduced by the greatest amount when compared with the whole maize, while flour (< or =500 microm) could be both reduced or increased depending on the mill and consignment. Thus, in most situations mycotoxin concentrations in whole maize that meet European Commission legislation on intake should give rise to levels in milled ingredients that should also do so. However, this was not always true in some ingredients, especially for fumonisins in those fractions with particle size < or =500 microm.

Occurrence and fate of Fusarium mycotoxins during commercial processing of oats in the UK.

The commercial processing of oats is different from that of other cereals, such as wheat and maize. In northwest Europe, oats also appear to be more susceptible to contamination with HT-2 and T-2 toxins than other cereals. Mycotoxins, such as deoxynivanol and zearalenone, in cereals are already controlled by EU legislation. With regard to additional, impending legislation, this study examined HT-2 and T-2 toxins together with zearalenone, deoxynivalenol and other related toxins in a commercial oat mill and how the concentrations varied from raw oats to the final prepared oat flakes. Concentrations of each Fusarium mycotoxin fell by 90-95% during the process, with the major loss being a physical distribution occurring at the de-hulling stage. Initial studies of losses occurring at other stages, such as kilning or de-branning of prepared oat groats, suggest these to be small. The use of colour sorting after kilning showed higher concentrations of each mycotoxin in the discoloured groats. The feasibility of developing a predictive tool for the oat industry is examined.

Fate of ochratoxin A in the processing of whole wheat grain during extrusion.

Ochratoxin A in cereal foods and some other products is controlled in the European Community by recent legislation with the objective of minimizing consumer exposure to this mycotoxin. Few studies have examined losses during processing. The stability of ochratoxin A during extrusion of contaminated wholemeal wheat flour was examined using pilot scale equipment. Factors examined were temperature, moisture content, screw speed and residence time. Ochratoxin A was partially stable with breakdown, increasing with temperature and moisture content. However, even under the harshest conditions likely to be used in commercial practice, maximum loss was no greater than 40%, with a residence time of about 40 s. The chemical properties of ochratoxin A suggest that breakdown might be affected by changes in pH and that further studies are necessary to investigate this possibility.

Formation of ochratoxin A and aflatoxins following the use of propionic acid as a grain preservative for storing damp barley.

The growth of storage moulds was studied in barley at 22% and approximately 28% moisture content treated with the recommended and reduced commercial doses of propionic acid over a 6 month storage period at 20°C. Experimental sample size was 5 kg barley per lot. Barley was fully protected against the growth ofA. flavus and aflatoxin formation when the recommended dose was applied. However, the treatment was less effective in controlling growth ofP. verrucosum and preventing ochratoxin A formation such that by 4 to 6 months of storage, the fungus had started to develop and toxin had formed even in some of the samples treated with propionic acid. The risk of the development of ochratoxin A during storage increased as the optimum dose was reduced, particularly for barley at 22% moisture content.

Fate of ochratoxin A in the processing of whole wheat grains during milling and bread production.

Batches of whole wheat contaminated with ochratoxin A were produced by inoculation with Penicillium verrucosum under controlled conditions in the laboratory. The fate of ochratoxin was followed through initial cleaning, abrasive scouring of the outer grain coat, milling into wholemeal wheat or into 10 milled fractions. Bread was baked from both wholemeal flour and straight-run white flour. Concentrations of ochratoxin A in the cleanings, scourings, and the bran and offal fractions were increased, but reduced in the white flour. Scouring removed up to 44% of the ochratoxin A present, but only a small further loss occurred in the bread-making process. An overall reduction of about 75% could be achieved in white bread using a combination of cleaning scouring and removal of the bran and offal fractions. Maximum overall reduction in producing wholemeal bread was about 40%. The reduction in ochratoxin A that can be achieved must be considered in relation to economic constraints concerning the disposal of wasted grain. Appropriate strategies for the use or disposal of potentially highly contaminated cleanings, scourings, bran or offal must be established.

Survey for aflatoxins, ochratoxin A, zearalenone and fumonisins in maize imported into the United Kingdom.

This survey examined 140 samples of raw maize as received at ports or at major maize mills in the UK and 12 after initial cleaning. Samples were examined for aflatoxins B1, B2, G1 and G2, ochratoxin A, zearalenone and fumonisins B1, B2 and B3 using fully validated analytical HPLC methods with detection limits of 0.1 microgram/kg for each aflatoxin and ochratoxin A, 4 micrograms/kg for zearalenone and 10 micrograms/kg for each fumonisin. 95.0% and 92.1% of samples met the new EC statutory maximum permissible level for total aflatoxins and aflatoxin B1 respectively. The maximum concentration of ochratoxin A found was 1.5 micrograms/kg. Zearalenone and fumonisins were detected in almost every sample with 41.7% of maize containing more than 100 micrograms/kg of zearalenone and 48% of samples containing more than 1000 micrograms/kg total fumonisins. Initial cleaning of raw maize reduced aflatoxin concentrations by about 40% and total fumonisins by 32%.

Surveillance of stored grain from the 1997 harvest in the United Kingdom for ochratoxin A.

This survey examined 306 samples of farm-stored wheat, barley and oats as received at, or tested by, central grain depots in the UK. Samples were taken from lorries or from stored grain using the existing in-house procedures used for quality checking and examined for ochratoxin A using a fully validated analytical HPLC method with a detection limit of 0.1 microgram/kg. Ochratoxin A was detected in 21% of the samples examined, with barley more frequently contaminated than wheat. Mean concentrations of ochratoxin A found for all samples were 0.69 microgram/kg in barley, 0.29 microgram/kg in wheat and 0.15 microgram/kg in oats. The highest concentration found was 17.8 micrograms/kg in a barley feed although concentrations of 81 and 30 micrograms/kg were found in 'reject-grade' wheat samples whose results were excluded from the main survey. In summary, 2.7 and 0.3% of samples exceeded concentrations of 5 and 10 micrograms/kg respectively. There appeared to be significant relationships between ochratoxin A concentrations and moisture content, storage time and geographical area. Although conditions at harvest in 1997 were quite variable countrywide and often wet, results were similar to those found in earlier surveys carried out in the UK.

A collaborative study of an HPLC method for determination of ochratoxin A in wheat using immunoaffinity column clean-up.

Thirteen laboratories within the United Kingdom participated in a collaborative study to determine ochratoxin A in wheat using an HPLC method incorporating immunoaffinity column clean-up. Mean recovery of ochratoxin A from wheat spiked at a level of 5 micrograms/kg was 91% for the method when using the OchraTest and 93% for the OCHRAPREP brands of column. Four samples naturally contaminated with ochratoxin A consisting of two blind duplicates containing concentrations of approximately 3 and 6 micrograms/kg of ochratoxin A were examined using a specified method. The relative standard deviations obtained within laboratories (repeatability) ranged between 9.7 and 12.5%, there being no significant difference between the two brands of immunoaffinity column. The relative standard deviations obtained between laboratories (reproducibility) were larger, ranging from 24.6 to 32.1% when using the OchraTest column and 14.0 and 19.6% for the OCHRAPREP column. When results were corrected for recovery, the apparent difference in reproducibility between the two columns was much reduced, 14.0-17.3% for OchraTest and 10.4-17.5% for OCHRAPREP. Horrat values for reproducibility were between 0.4 and 0.7 before correcting results for recovery and were between 0.3 and 0.6 after correction. The performance of the immunoaffinity column method tested compared very favourably with results of other published collaborative studies on mycotoxins.

Mycotoxins in ingredients of animal feeding stuffs: III. Determination of mycotoxins in rice bran.

Methods used previously for the determination of aflatoxins B1, B2, G1 and G2, ochratoxins A and B, cyclopiazonic acid, zearalenone, sterigmatocystin, and moniliformin in maize were applied successfully to rice bran. However, recovery of deoxynivalenol and other trichothecene mycotoxins spiked into samples was lower than expected and no citrinin could be recovered. Forty samples of rice bran used in the animal feed industry were examined for the presence of 20 mycotoxins. The level of contamination of rice bran by mycotoxins was low. Aflatoxin B1 was present in 29 (72.5%) samples, usually together with related aflatoxins, up to a total of 28 micrograms/kg after allowing for recovery losses. Ochratoxin A at a level of 12 and 3 micrograms/kg was confirmed in two samples, while most samples also appeared to contain small concentrations of ochratoxin A. A chromatographic peak corresponding to cyclopiazonic acid occurred in several samples and one sample appeared to contain a low level of moniliformin. Positive confirmation of these last two mycotoxins was difficult and these findings must be regarded with caution. No other toxins were detected.

Mycotoxins in ingredients of animal feeding stuffs: II. Determination of mycotoxins in maize and maize products.

Analytical methods have been developed for the reliable detection and estimation of 22 mycotoxins in maize gluten and other maize products used in the animal feed industry. The mycotoxins are aflatoxins B1, B2, G1 and G2, ochratoxins A and B, citrinin, cyclopiazonic acid, zearalenone, sterigmatocystin, deoxynivalenol, nivalenol together with seven related trichothecene mycotoxins, fumonisins B1 and B2 and moniliformin. For most of the mycotoxins, recoveries obtained were 60% or greater and reproducibility data were better than +/- 40%. In general, the analysis of maize gluten proved more difficult than for other maize products. In total 40 samples of maize gluten, and 27 samples of other maize products were examined. Aflatoxins were not found above the reporting limit (1-5 micrograms/kg) in any sample while ochratoxin A was detected in only two samples of maize gluten at 2 micrograms/kg. No sterigmatocystin or cyclopiazonic acid were detected although the limits of detection for these toxins were poor. Twenty percent of maize gluten samples contained zearalenone at up to 500 micrograms/kg while all other maize products also contained this mycotoxin. Highest levels occurred in screenings and meal while the lowest amounts were in flaked maize and germ. Many samples contained a multi-toxin mixture of trichothecenes, fumonisins and moniliformin. The most highly contaminated samples of maize screenings and maize meal contained a mixture of zearalenone, deoxynivalenol, nivalenol, fumonisin B1 and B2, moniliformin, each in mg/kg amounts and lower amounts of other trichothecenes such as 15-acetoxy deoxynivalenol, HT-2 toxin and T-2 toxin. Fumonisins occurred in all except two gluten samples and occurred up to a level of 32 mg/kg in maize screenings, 13 mg/kg in maize meal and 8 mg/kg in maize germ.

Mycotoxins in ingredients of animal feeding stuffs: I. Determination of Alternaria mycotoxins in oilseed rape meal and sunflower seed meal.

A multi-toxin method was developed for the detection of some of the known Alternaria mycotoxins, altenuene, iso-altenuene, alternariol, alternariol monomethyl ether, tenuazonic acid and altertoxin I in oilseed rape meal and sunflower seed meal. The method involves extraction of the toxins with an acidified mixture of acetonitrile: aqueous potassium chloride solution, followed by liquid-liquid extraction and further purification using gel permeation chromatography. The final extract is then examined on a reverse phase high performance liquid chromatographic gradient system with both fluorescence and UV detection. The average recoveries found were 94, 84, 109, 85, 66 and 93% for spiked oilseed rape meal samples and 91, 89, 96, 75, 61 and 102% for spiked sunflower meal samples with limits of determination of about 40, 50, 50, 40, 350 and 200 micrograms/kg for the above toxins, respectively. Detection limits were about 30% of these values. Thirty samples of oilseed rape meal and 22 samples of sunflower meal were examined using the methods developed. Twenty of the oilseed rape products which had been grown in the UK were free from contamination while 10 contained one or more of tenuazonic acid, alternariol and alternariol monomethyl ether. In contrast, all of the sunflower meal samples, of Argentinean, Indian or EC origin, were contaminated with one or more of alternariol, alternariol monomethyl ether and tenuazonic acid. Average levels of alternariol, alternariol monomethyl ether and tenuazonic acid were 68, 55 and 730 micrograms/kg, respectively for the contaminated samples of oilseed rape meal and 180, 100 and 1900 micrograms/kg, respectively for the contaminated samples of sunflower seed meal.

Occurrence of mycotoxins in raw ingredients used for animal feeding stuffs in the United Kingdom in 1992.

Examination of 330 samples of animal feed ingredients for the presence of a number of mycotoxins has been carried out. These samples were drawn from 186 animal feed mills in the United Kingdom. Aflatoxin B1 was the mycotoxin found most frequently, occurring in most samples of rice bran, maize products, palm kernels and cottonseed, but not in only 3 out of 20 samples of sunflower, in 1 out of 20 samples of soya and in no samples of peas, beans or manioc. Analytical difficulties were met with some combinations of commodity and mycotoxin and all results are uncorrected for recovery. The highest level was detected in a sample of maize gluten: 41 micrograms/kg of aflatoxin B1 (47 micrograms/kg total aflatoxins). Maize products also frequently contained fumonisins B1 and B2 at levels up to nearly 5,000 micrograms/kg in total and zearalenone up to a maximum level of 500 micrograms/kg. Ochratoxin A and citrinin were found in approximately 20% of wheat and barley samples. One sample of barley contained ochratoxin A at a level of 102 micrograms/kg and citrinin at a level of 8 micrograms/kg. Low levels of ochratoxin A also occurred in a few samples of other ingredients: rice bran, palm kernel and beans. Sterigmatocystin at 18 micrograms/kg was found in one sample of organically grown wheat and a trace amount of zearalenone in one sample of manioc. Multi-mycotoxin contamination also occurred, particularly in some samples of maize for which 19 out of 50 samples contained both aflatoxins and fumonisins.

Determination of mycotoxins in pet foods sold for domestic pets and wild birds using linked-column immunoassay clean-up and HPLC.

A technique referred to as 'linked-column immunoassay clean-up', was developed for the simultaneous determination of aflatoxins and ochratoxin A in a range of dry cereal-based pet foods and wild bird food. In addition a method also based on clean-up using immunoaffinity columns was used for the determination of fumonisin mycotoxins in samples known or suspected to contain maize. One hundred samples of pet foods consisting of 35 samples of domestic bird seeds and 15 samples of wild bird food were examined for aflatoxins B1, B2, G1 and G2 and ochratoxin A. Twenty samples of these samples were also examined for fumonisins B1 and B2. Limits of detection were about 0.5 micrograms/kg for each aflatoxin and ochratoxin A and 3 and 8 micrograms/kg for fumonisins B1 and B2 respectively. Eighty-four percent of the samples contained no measurable concentrations of mycotoxins. A low level of aflatoxin B1 was found in a sample of cat food and a concentration of 370 micrograms/kg aflatoxin B1 in one sample of peanuts marketed for wild birds. Ochratoxin A was detected in 10% of samples but in low concentrations, the highest of 7 micrograms/kg occurring in a sample of bird food. Fumonisins were found in 30% of the 20 samples tested with a maximum of 750 micrograms/kg total fumonisins being found in a sample of cat food. Five samples each of dog and cat foods were examined for mould count and fungal species as received and after storage under controlled simulated damp conditions. Mould counts in all 10 samples examined were very low when received. The samples in which moisture content had been increased contained visible mould after storage. Penicillium, Eurotium and Aspergillus were the predominant species. However, no ochratoxin A or aflatoxins were detected in these spoiled samples.

Analytical methods for the determination of sterigmatocystin in cheese, bread and corn products using HPLC with atmospheric pressure ionization mass spectrometric detection.

Methods have been developed for the determination of sterigmatocystin in bread, maize and cheese using HPLC linked to mass spectrometry (MS) atmospheric pressure chemical ionization for detection and concurrent confirmation of sterigmatocystin at levels down to less than 5 micrograms/kg. Candidate extraction methods were initially checked for recovery and reproducibility by spiking commodities at a level of 200 micrograms/kg of sterigmatocystin and using HPLC with post-column derivatization. Recovery was found to be greater than 90% for bread and maize, and 75% for cheese. Mass spectrometer conditions for detecting sterigmatocystin were established by injecting solutions directly into the mass spectrometer. Extracts of bread, maize grits and cheese prepared by the candidate extraction methods were examined using HPLC/MS using samples spiked at a level of 20 micrograms/kg of sterigmatocystin. Results for bread and maize samples showed that the extraction procedure recovered more than 90% of added sterigmatocystin and produced extracts free of interference from co-extractives, with limits of detection of less than 2 micrograms/kg for both commodities. The HPLC/MS results for cheese extracts gave lower average recoveries of 55%. These results were also more variable. However, the apparent limit of detection for sterigmatocystin in cheese was still about 4 micrograms/kg.

The effects of milling and processing on wheat contaminated with ochratoxin A.

Samples of sound home-grown wheat (one hard and one soft milling) were obtained, cleaned, and gamma-irradiation used to reduce numbers of viable naturally-occurring fungi. Each sample was inoculated with a toxigenic strain of Penicillium verrucosum and monitored for ochratoxin A formation. When ochratoxin A had reached a level of 60 micrograms/kg, the samples were milled into ten fractions which were analysed for ochratoxin A by an HPLC method with immunoaffinity column clean-up. Each straight-run white flour was baked into bread which was analysed in the same way. Relationships between ochratoxin A levels in naturally-contaminated wheat and the products of milling and baking were established. The recovery of ochratoxin A in wholemeal compared with the cleaned wheat was essentially complete and no significant loss occurred on baking white or wholemeal flour into bread. Recoveries in the straight-run white flours, however, were only approximately one-third for the hard wheat and two-thirds for the soft wheat of the ochratoxin A in the uncleaned wheat. The reason for this was that a much higher proportion of the ochratoxin A was found in the bran and offal fractions from hard wheat than from soft. Conversely, a much higher proportion of the ochratoxin A was found in the reduction flour from soft wheat than from hard. Scouring was examined as a possible method of decontamination of wheat prior to milling. This process removes a proportion of the pericarp (bran coat) prior to milling. The results of the study confirmed that scouring reduced the ochratoxin A level in white and wholemeal flour three-fold for both the hard and soft wheat.

Ochratoxin A in animal feed--effects of processing.

The presence of ochratoxin A in ingredients of animal feeding stuffs is unlikely to produce noticeable effects on animal health or productivity except in the unlikely event of gross contamination. Studies on ochratoxin in cereals and food products have suggested that this mycotoxin is relatively stable under a range of conditions. This raises the question of whether ochratoxin can occur in meat and other animal products, and if so, to what extent this contamination contributes to the human diet. This paper reviews the information on the likely fate of ochratoxin at each stage from raw product to human food. It concludes that small amounts of ochratoxin may survive processing and animal metabolism in products from pigs and poultry but it is highly unlikely to be found in milk or meat from cattle.

Natural occurrence of mycotoxins and mycotoxigenic fungi in cereals in the United Kingdom.

Approximately 50 samples of poorly stored cereals were examined for mycotoxins, fungi and biological activity. While some mycotoxins such as ochratoxin A and citrinin occurred frequently, many other fungal metabolites present could not be identified. Cereal extracts and isolates of some fungi, when grown in culture, were often toxic to brine shrimp larvae but, in some toxic extracts, no known mycotoxins were identified. A preliminary report of this study forms the subject of this paper.

Detection of zearalenone in cereal extracts using high-performance liquid chromatography with post-column derivatization.

Post-column derivatization has been used to enhance the fluorescence response of the Fusarium mycotoxins zearalenone and zearalenol when determined by reversed-phase high-performance liquid chromatography. Derivatization is based on reaction with aluminum chloride and this results in a more selective response for these toxins. The method was tested on a number of cereals and animal feeds.