Prevention of ochratoxin A in commodities and likely effects of processing fractionation and animal feeds

Abstract

Whenever possible, preventing the formation of ochratoxin A (OTA) in susceptible food commodities such as cereals, grapes and coffee beans should be the primary aim. Each product tends to host a specific OTA-producing mould so that the environmental conditions and factors that encourage the subsequent formation of OTA need to be understood. Codes of Practice for prevention and management of OTA are being developed and can be used in conjunction with a HACCP approach to protect the end consumer, in line with EU statutory limits. If prevention fails, an understanding of how concentrations change during the whole food chain may be useful in minimizing the concentrations reaching the consumer. OTA is quite heat stable under neutral conditions but may be partly broken down, e.g. in extrusion processing. In milling or other separation procedures, OTA will be concentrated or reduced in the resulting components. By-products such as ‘cleanings’ or bran may contain high concentrations and are often used for animal feed. Introduction of guideline or statutory maximum concentrations for feed within the EU makes it essential that concentrations of ochratoxin A in such by-products are acceptable. This paper reviews recent literature and findings from recent 5th Framework EU-funded projects.

Keywords:

• OTA
• prevention
• processing
• cereals
• milling
• bread
• extrusion
• grapes
• coffee
• animal feed

Introduction

Ochratoxin A (OTA) is a kidney toxin and a possible genotoxic compound. It occurs in a range of food crops including cereals, coffee, grapes, cocoa beans and pulses (European Commission 2002), usually developing after harvest due to storage problems (Lacey & Magan 1991). Historically the literature records many fungal species that appeared to form this mycotoxin but many of these have been shown not to be true OTA producers but were misidentified because of taxonomic difficulties or isolation from impure cultures (Frisvad 1989). It is well established that cereals, particularly in Northern and Western Europe, Canada and other temperate areas, are at risk from the formation of OTA originating from Penicillium spp and provide the largest contribution of this mycotoxin to human intake. Coffee beans, grapes, cocoa beans and spices require warmer conditions for development and OTA contamination of these is due mostly to Aspergillus spp.

The EU has recently introduced maximum permissible limits for OTA to reduce risk to the consumer. In cereals, these limits are 5 µg kg⁻¹ for whole grain and 3 µg kg⁻¹ for processed products.

The need to limit OTA intake has led to an EU Research programme. In project (OTA-PREV, 2000–2003) funded under the 5th Framework Programme many aspects of the formation, prevention and fate of OTA in cereals were studied. Some of the results of this study are presented within this review. A second study funded under the same Framework (WINE-OCHRA RISK, 2001–2004) addresses the problem of OTA in grapes and wine while the European Coffee Federation (www.ecf-coffee.org) has been active in studying the formation and fate of OTA in coffee beans.

Findings from these studies have contributed to the development of Codes of Practice for the prevention of ochratoxin in cereals, grapes and coffee beans based on Good Agricultural Practice (GAP) and Good Manufacturing Practice (GAM). Management systems for these products based on Hazard Analysis Critical Control Points (HACCP) are being developed.

The analysis, surveillance, evaluation and control of OTA in food have been previously reviewed by ILSI (Battaglia et al. 1996). This current paper reviews recent findings on the prevention of OTA with particular emphasis on cereals and means for its management in the processing chain.

Prevention of OTA formation in cereals

The principles of good storage practice have been known for many years and if followed should prevent the formation of OTA and other storage mycotoxins after harvest. These principles are governed by two basic concepts. The first is that the infection of the cereal by OTA-toxigenic moulds must be prevented or reduced as far as is possible. The second arises because it is impossible to completely eliminate all sources of mould infection so it is then important to avoid conditions conducive to mould growth. In simple terms, the cereal must be dried below 15% (a water activity of about 0.8) and kept at this level throughout storage. However, OTA continues to be a problem in cereals both in developing and developed countries. In studies in the field, OTA has usually, but not always, been absent in cereals at harvest (Elmholt 2003). When it has been detected, this is most likely because insufficient care has been taken to prevent mould growth in the sample between collection and analysis. It has been shown that OTA in cereals is principally produced after infection with Penicillium verrucosum (Frisvad & Lund 1993) and originates because of poor storage practice.

During the OTA-PREV study, only P. verrucosum was isolated from OTA positive cereal samples obtained within Europe. In the course of that study a questionnaire was sent to cereal and storage experts in most EU Member and a few other States concerning many aspects relating to OTA. While the replies did not reflect the official view of each State, OTA was not considered by industry to be a major problem other than due to the requirement to meet statutory maximum permitted levels. Replies confirmed that grain (wheat, barley, maize and oats) is sometimes harvested at high moisture contents of 20% and above, and drying to safe storage moisture content is then universally recognized as necessary to prevent the subsequent development of storage moulds.

It was clear, however, from the replies to the questionnaire that sometimes grain, is not, or cannot be dried quickly enough at harvest. Delays in drying then put the grain at risk that can lead to subsequent problems during storage. This is particularly a problem in those regions where the rainfall during the harvest is unpredictable and economic considerations do not allow expensive drying machinery to lie idle for a number of years when it is not required. Countries that expect rainfall during the harvest period are usually better prepared with efficient drying machinery and this helps to avoid storage problems. The practice of on-floor ambient drying used in some Community States has the inherent potential for problems to occur when atmospheric conditions prevent rapid drying throughout the grain bulk. Grain in flat stores or bins is dried from the bottom upwards so that the upper layers remain at risk for much longer and this method further depends on effective maintenance of the dryers and air channels.

Failure to dry cereals sufficiently for safe long-term storage is thus the main reason why OTA continues to occur. ‘Moisture content’ is normally used for measurement and decision-making in the practical situation but ‘water activity’ is a more accurate measure of the water present in a commodity and is used by mycologists to indicate the rage over which P. verrucosum can grow and form OTA. The actual amount of OTA that occurs and the time necessary for this in any situation will depend on a number of key factors. These are a source of infection with P. verrucosum, water activity rising above a critical value (Magan 2005), temperature and the storage time. Many other factors will affect the final mycotoxin level. Table I lists some of these factors that are often interdependent. Cereals that suffer stress in the field during seed development are likely to be more easily colonized by fungi. Thus for example, P. verrucosum present in harvesters and trailers gains easier access to the grain structure. Then if climatic conditions are wet, it becomes even more critical to dry to safe storage water activity as quickly as possible to avoid the development of OTA.

Table I. Factors affecting cereals at harvest that may influence the formation and concentrations of OTA.

Factors
Weather before and at harvest
Time before drying
Efficiency of drying machinery
Physical state of grains
Temperature at harvest
Fungal competition
Cleanliness of harvesters and transport

Once in store, cereals must be regularly inspected and kept under safe storage conditions. Some important factors for keeping grain safe are given in Table II. Hygiene and thorough cleaning of stores, silos and transporting containers and trucks are crucial to minimize the residual sources of inoculum (P. verrucosum). In a UK study, approximately 40 samples of sweepings from the floor and grain bins, residues of grain left in bins and other waste grain were collected from a number of typical grain stores or farms. All samples contained detectable P. verrucosum, some levels as high as log 11 colony forming units g⁻¹, and most contained detectable concentrations of OTA. The results for 5 typical samples are given in Table III. These samples highlight a double hazard; that the fungal spores are a source of inoculum for future grain deliveries and that the return of such waste grain to the original bulk could contaminate a considerable weight of clean grain with OTA and represents a practice that must be strongly discouraged.

Table II. Factors affecting cereals in store that may influence the formation and concentrations of OTA.

Factors
Cleanliness of storage containers
Absence of structural leaks
Condensation
Temperature
Time

Table III. P. verrucosum and OTA in residual grain and dust collected from grain stores in the UK.

Selected samples (kg)	P. verrucosum, log cfu g^−1	OTA, µg kg^−1
Barley and dust (100)	3.7	60
Dust and cereal debris	4.7	615
Barley (2)	5.3	133
Wheat debris	8.2	437
Mouldy wheat (20)	8.6	>1000

Codes of Practice exist or are being developed for the safe storage of grain. When the OTA-PREV questionnaire was conducted it appeared that there were approximately 6–7 million farms within the 15 Member States most of whom produced or stored cereals. The logistics of effective dissemination of information on safe storage of grain was recognized as a major problem for some countries where the infrastructure may be such that sufficient funds and expertise are unavailable to advise on and ensure best storage practice.

While this appraisal addresses cereals here, similar problems and principles apply to other key commodities such as grapes, coffee beans and cocoa beans also at risk from OTA. Parallel studies have been or should be carried out in order to develop practical Codes of Practice and guidelines in conjunction with the relevant industries.

Processing and management of OTA in the food chain

Effects of processing

Preventing the formation of OTA should be the primary objective of any mycotoxin management system and in many instances can be achieved if a few basic rules are followed. Drying and subsequent storage of grain under ‘safe’ conditions should reduce or eliminate the risk of occurrence of OTA and this has been known for many years. However in practice, this is difficult to achieve and surveys continue to find contaminated grain and cereal food products (e.g. Scudamore et al. 1999, Wolff 2000, European Commission 2002). Thus, when prevention proves impossible, a thorough understanding of the effects of ‘processing’ in all its forms can often be used to minimize the exposure of the end consumer, whether human or animal.

A working definition of ‘processing’ has been offered as “the application of any combination of chemical, biological or physical methods used to produce the final consumer food or animal feed” (Scudamore 2005). There are two principle means by which OTA concentrations can change during processing. The first stems from physical separation such as cleaning, removal of husks, seed coats and shells, milling into component fractions and partition between the solid grain phase and liquid when steeping in aqueous solution. The second results from chemical or enzymic reactions that are used in commercial food or feed production. A number of studies, some cited below, have attempted to quantify the conditions and the extent under which OTA is degraded during the preparation of cereal products such as bread and breakfast cereals or products from other key commodities.

A summary of the main operations that can affect concentrations of OTA in the food chain is given in Table IV. Most operations tend to reduce concentrations or have no effect in the primary product although increased concentrations may occur in the by-products such as cleanings, surface scourings or bran products.

Table IV. Operations, conditions and effects that may affect concentrations of OTA in cereals.

Operations	Effects	Change in OTA
Cleaning and aspiration	Removes dust, fungal spores	−
Scouring	Takes of outer layers	−
Grading/sorting	Removes low quality grains	−
Milling	Grains separated into fractions	− and +
Steeping	Mycotoxin dissolves in aqueous solution	− or 0
Other ingredients	May modify chemical reactions	− or 0
Temperature	Reaction increases with temperature	−
pH	Effects reactivity	− or 0
Moisture	Increases hydrolysis	− or 0
Pressure and time	Breakdown will increase with time	− or 0
Enzyme action	May hydrolyse	− or 0
Residual mould	In situ formation of OTA	+
Ingestion of animal feed	Transfer to animal tissues and blood	−
OTA level in primary product: − = reduction, 0 = no change, + = increase.

Removing the most contaminated elements will assist in reducing the exposure of the end consumer to OTA. The use of these by-products for animal feed needs to be limited to concentrations that are consistent with acceptable levels of carry over to meat and other animal products and/or significant effects on the productivity or health of the target livestock or poultry.

Stability of OTA

Relatively few studies have been carried out on the chemical breakdown of OTA during processing although it is considered to be a relatively stable compound when heated under neutral conditions. The fate and factors that affect levels of mycotoxins including OTA during processing have been reviewed (Scudamore 2005).

Boudra et al. (1995) studied the decomposition of OTA under different moisture content and temperature. Their results showed that OTA could be partially decomposed so that in the presence of 50% water, loss in comparison to drier grain is increased at 100 and 150 but is decreased at 200°C.

Cooking of polished wheat using a procedure common in Egypt only removed 6% of OTA (El-Banna & Scott 1984) and a similar result was demonstrated in that when whole both hard and soft wheat containing OTA at about 60 µg kg⁻¹ were milled white flour contained reduced amounts of OTA although subsequently only a small further reduction occurred when this was baked into bread (Osborne et al. 1996). During a previous Workshop sponsored by ILSI (Battaglia et al. 1996) various workers reviewed specific aspects of the problem; the effects of processing and detoxification treatments on OTA (Scott 1996), the fate of OTA during bread making (Subirade 1996), the effects of processing on the occurrence of OTA in cereals (Alldrick 1996) and fate during malting and brewing (Baxter 1996). OTA is thus relatively stable once formed, but under certain conditions of high temperatures and acid or alkaline conditions or in the presence of enzymes, some breakdown can occur. OTA in neutral conditions is relatively stable during boiling processes although citrinin that is often present together with OTA readily decomposes (Jackson & Ciegler 1978, Madsen et al. 1983).

Considerable care must be observed when comparing experimental laboratory results with those obtained in commercial practice. The exact recipes and processing conditions used by industry are often of a highly sensitive nature and are rarely made public. For example, breakfast cereals such as wheat products or cornflakes can be produced both by cooking and by extrusion. This can involve cooking under pressure and boiling temperatures with undisclosed recipes. When experimental data has been obtained from extrusion for example, full details of the physical conditions are often not presented. Temperature, pressure, moisture content, screw configuration and speed may all contribute to any chemical breakdown but complete data are rarely given. If the time within the extruder is not stated only limited conclusions about the breakdown can be made because chemical reaction is dependent on extruder residence time.

Milling

OTA tends to be concentrated in the outer bran layers of cereals (Osborne 1979, Osborne et al. 1996) and this raises the question of redistribution during milling so that both reduction and increase in concentration can occur depending on the milled fraction examined. In contrast, studies in Poland showed that cleaning and milling wheat and barley did not remove OTA in naturally contaminated samples, and levels in flour and bran were the same (Chelkowski et al. 1981).

Studies within OTA-PREV were designed to confirm and clarify these earlier results. The aim was to establish the change in concentration of OTA during milling, abrasive scouring of the grain and baking and to assess the distribution of OTA in all milled fractions. The methodology and results have been presented in detail (Scudamore et al. 2003). Contaminated baking quality wheat, variety Hereward, was prepared by inoculation of 100 kg batches of grain of 20% moisture content with the causative fungus (P. verrucosum) under carefully controlled conditions. The wheat was held at 20°C and samples withdrawn and tested for concentration and homogeneity until levels reached approximately 5 µg kg⁻¹ or 50 µg kg⁻¹ of OTA. After drying to safe moisture content, the wheat was cleaned, conditioned and milled in a pilot-scale mill, mimicking commercial practice.

Milling was used to either produce wholemeal flour or to separate wheat into component fractions that comprised three reduction flour portions, three break flour portions, bran, bran flour, offal and offal flour fractions. The definition of these fractions is provided in the reference. In addition, before milling the whole wheat, batches were divided equally into three portions. One of these was left untouched before milling but the other two were subjected to abrasive ‘scouring’ that aimed to remove about 1% and 2% by weight of the outer seed coat layers. All cleanings, fractions and scourings were analysed.

Bread production and extrusion of wholemeal wheat flour

The fate of OTA during bread making has been reviewed previously (Subirade 1996) while baking bread from test flour showed that there was little loss during this stage (Osborne et al. 1996). This was further studied in OTA-PREV (Scudamore et al. 2003) using milled wholemeal and white flour prepared as described above for baking into bread using a pilot scale bakery that replicated the ‘Chorleywood Bread Process’. This procedure was developed in 1961 by cereal scientists (Chamberlain et al. 1962, Axford et al. 1963) and widely adopted by industry.

The wholemeal wheat containing OTA prepared during OTA-PEV Project was also used for studying the effect of extrusion and the work has also been published (Scudamore et al. 2004). Wholemeal flour was fed through the extruder using residence times between 25 and 50 seconds. Different screw configurations, water content and temperatures were used and samples from each set of conditions were analysed.

Results of milling baking and extrusion from an EU-funded Project, OTA-PREV

These results have been discussed in detail (Scudamore et al. 2003, 2004) but are summarized here. Figure 1 shows the changes in OTA concentration during the whole process that produces bread. The left hand histogram is for white bread from white flour while that on the right is for wholemeal flour and bread. Initial cleaning results in removing a small amount of material so that only a small drop in OTA in the whole wheat. Even so, the corresponding concentrations in the screenings were as high as 50 µg kg⁻¹.

Figure 1. The change in OTA concentrations during, cleaning, scouring, milling and bread making, starting level 6 µg kg⁻¹.

Column set three shows the effect of removing 1 and 2% by weight of the surface layers of wheat and that this effectively reduces levels in the cleaned wheat by a further 25–40%. Again the concentrations in the scourings were much higher, at approximately 100 µg kg⁻¹. When the scoured and unscoured wheat lots were milled there was no loss in wholemeal flour but OTA was much reduced in the white flour samples while being concentrated in the bran and offal portions. In the last bread-baking step, no further loss or change in OTA concentrations occurred. Thus overall about 30% was lost from intake wheat to the wholemeal bread, while up to 75% of OTA was lost from wheat to white bread. This is more or less consistent with the EU 2-tier legislation for OTA in cereals of 5 and 3 µg kg⁻¹, respectively for raw cereal and finished product.

Figure 2 shows the concentration of OTA in the 10 common fractions obtained by milling. In this batch a high starting level of OTA was used and this showed an approximate concentration of x3 in the bran. This proportional increase was similar to that obtained in other samples containing a starting concentration close to the EU limit of 5 µg kg⁻¹. It is thus clear that bran-based products processed from wheat at the EU limit of 5 µg kg⁻¹ are likely to contain concentrations of OTA perhaps as much as x5 the EU maximum permissible limit of 3 µg kg⁻¹.

Figure 2. Distribution of OTA of a concentration of 42.2 µg kg⁻¹ in whole wheat in milled wheat fractions.

Figure 3 summarizes the results obtained during extrusion at different temperatures for the wholemeal wheat samples between 17 and 20% water content which shows a clear relationship with temperature. OTA content was corrected for changes in moisture content within the extruder and in the extruded product. However, when water content was about 30% (see reference) this appeared to accelerate breakdown at lower temperatures but to result in a similar breakdown to that achieved with 17/20% wheat at the higher temperatures. This seems to support the findings of Boudra et al. (1995) who suggested a much accelerated breakdown of OTA in high moisture wheat, but only at lower temperatures. Most commercial extrusion processes are not operated above about 180°C so this study suggests that losses during extrusion under these conditions is limited to about 25%. However, different recipes or low or high pH conditions may stabilize or accelerate breakdown. This study did not consider the nature of the breakdown seen.

Figure 3. Breakdown of ochratoxin A during extrusion processing, moisture content in the range 17–20%.

Fate of OTA during brewing

Scott (1996) showed that by adding OTA at various stages during the brewing process it could be transmitted into beer to some extent. About 2–13% of OTA is destroyed during the fermentation stage (Scott et al. 1995).

The fate of OTA during malting and brewing has been reviewed (Baxter 1996). Malt containing OTA at about 50 µg kg⁻¹ produced by inoculation with P. verrucosum was used in a pilot plant to determine the fate of OTA (Baxter et al. 2001). Overall between 13 and 32% of the OTA survived in the beer. Up to 40% was lost during mashing and another 16% remained in the spent grains.

A study of the brewing process within the laboratory and on a pilot-scale was carried out within the OTA-PREV project. An increase of 2–4-fold in OTA levels during malting was found in 75% of the samples studied. However, the process temperature had a pronounced effect so that at 16–18°C OTA formation was 20-fold compared to 5-fold between 12–14°C. During the complete brewing process approximately 20% of the original OTA from the malt remained in the beer.

Fate of OTA in other food commodities

OTA often occurs in green coffee beans so considerable research has been carried out to determine its fate during processing. Green coffee is cleaned, roasted, typically for 20 min at about 200°C, ground, and then extracted with hot water, concentrated and spray-dried. The fate of OTA during processing of coffee was reviewed (Viani 1996). The extent of loss reported in the literature varies considerably. Studer-Rohr et al. (1994) found OTA in 13 out of 25 samples of green beans analysed and the mycotoxin level was only slightly reduced after roasting, with most of the mycotoxin then eluting into the brew. Van der Stegen et al. (2001) also studied the effect of roasting conditions on reduction of OTA in coffee. A commercial lot of green coffee, naturally contaminated with OTA, was roasted under various conditions and the effects on its final OTA content were determined. Roasting time varied from 2.5–10 min and the roast colour varied from light medium to dark. The reduction was about 69% reduction over the combined results. Three different explanations are proposed for this reduction: Physical removal of OTA with chaff, isomerization at the C-3 position into another diastereomer, and thermal degradation with possible involvement of moisture. The authors suggest that all three possibilities may play a role in OTA reduction during coffee roasting. Blanc et al. (1998) investigated the fate of OTA along an industrial coffee manufacturing line. In naturally contaminated green Robusta coffee beans the OTA levels were drastically reduced during soluble coffee production. Some OTA was removed by initial cleaning of the beans but the main reduction occurred during roasting. The roast and the ground coffee contained only 16% of the starting concentration. A further 3% loss occurred in producing soluble coffee powder. Leoni et al. (2000) showed that between about 50–100% of OTA could be transferred from roast and ground coffee into the coffee brew.

Milanez and Leitao (1996) showed up to 84% loss of OTA in processing beans (Phaseolus vulgaris L.) although smaller losses of about 53% were reported previously after bleaching, salting and heat processing (Harwig et al. 1974). Greater losses were observed when beans were soaked in water for 12 h before cooking under pressure at 115°C for 45 min.

In a study to investigate the influence of the shelling process on the presence of OTA in 22 cocoa samples, 14 cocoa bean samples (64%) suffered a loss of OTA of more than 95% due to shelling, six samples suffered a loss of OTA in the range 65–95%, and only one sample presented a reduction of less than 50% in cocoa (Amézqueta et al. 2005). Here the principal conclusion was that OTA contamination of beans is concentrated in the shell and therefore improvements of the industrial shelling process could prevent or reduce OTA occurrence in cocoa final products.

OTA in animal feed

The risks posed by mycotoxins such as OTA in animal feed are the potential effects on animal health and productivity and the contribution to human exposure through consuming meat and other animal products. Exposure of the consumer to OTA by this means is considered to be very small and mainly restricted to pork and products based on pig's blood and organs. However, the EU is considering maximum permissible limits for a range of mycotoxins, including OTA, in animal feed and the appropriate values for these. The introduction of statutory or guideline limits for OTA poses potential problems for the feed industry as cereal by-products surplus to human food production are often key components of animal feed.

Material such as cereal cleanings, broken grains, wheat and maize bran, maize germ, gluten and meal, oat husks and feed barley are all products in which mycotoxins, including OTA, are often concentrated as discussed earlier. These may be fed directly or may be processed into compound feedstuffs by pelleting using heating, extrusion or other temperature or pressure based procedures. The fate of mycotoxins in animal feed production remains poorly studied with little further work reported since the review by Scudamore in 1996. The difficulty of sampling animal feed has been reviewed (Whitaker et al. 2005)

A summary of the results for OTA in which a range of mycotoxins were sought in more than 350 samples of ingredients used for animal feed in the UK (Scudamore et al. 1997) is given in Table V. These samples were supplied by commercial feed mills. It is unclear whether supplies included cereal by-products from the grain millers. These results confirm that cereals were the materials in which OTA occurred most frequently although some samples of rice bran, palm kernel meal, dried peas and beans were also positive. The incidence and concentrations of OTA in maize products in this work was however very low and is in accordance with samples taken in a survey of European and Argentinean maize imported into the UK (Scudamore & Patel 2000). However, OTA can occur in high concentrations in maize in some parts of Europe (Speijers & van Egmond 1993).

Table V. OTA in animal feed ingredients UK 1992 (Scudamore et al. 1997), Limit of quantification 1 µg kg⁻¹.

Ingredients	Number of samples	% Positive	Maximum value, µg kg^−1
Wheat	50	20	17
Barley	45	27	102
Maize fractions	50	0	0
Rice bran	40	8	6
Cottonseed meal	21	0	0
Rapeseed meal	25	0	0
Sunflower meal	20	0	0
Olive pulp	5	0	0
Palm products	15	13	6
Soya meal	20	0	0
Dried peas/beans	15	7	33
Manioc	10	0	0

Within the UK and other EU Member States, it is common for barley and wheat to be grown and stored on farm and fed directly to livestock without entering the cereal market. This grain is thus not traded or tested and little is known about the concentrations of OTA that may be present in these circumstances. It can only be surmised that the occurrence and amounts of OTA will be sporadic and variable and dependent on the hygiene and management protocols operated. Productivity or health problems are similarly likely to be intermittent and unlikely to be recognized as mycotoxin-related incidents.

Means of controlling mycotoxins in animal feed have been reviewed (Pettersson 2004). Studies into feeding mineral clays to counteract mycotoxins have shown that these have little beneficial effect for OTA (Santin et al. 2002). Recently, yeast cell wall-based adsorbents have shown considerable promise in their ability to remove several mycotoxins from feed ingested and this has been reviewed (Devegowda & Murthy 2005).

Summary and conclusions

Studies have emphasized that OTA in cereals is a storage problem that can be prevented if sound Codes of Practice are carefully followed and can be managed using a HACCP approach. Similar guidelines are being developed for other key food commodities such as grapes and coffee. Prevention should be the primary aim because OTA is quite heat stable and difficult to remove although it can be reduced at higher temperatures and during extrusion. OTA may increase during malting or may be reduced by enzymic action.

In wheat and maize milling OTA is concentrated in cleanings, bran and other fractions derived from the seed coat. However these by-products are often used for animal feed and maximum permissible guideline limits for OTA are likely to operate within the EU from January 2006. In addition, barley and wheat may be grown on farm and used directly for animal feed. Bran-based human foods present a particular problem in relationship to the 2-tier regulations of 5 and 3 µg kg⁻¹ for whole and finished cereal products because OTA is concentrated by a factor of approximately 3 on milling and is quite stable during most processes.

Acknowledgements

The Projects OTA-PREV and WINE-OCHRA RISK were funded under the EU 5th Framework Programme.