lable at ScienceDirect

Food Control 22 (2011) 1145e1153

Contents lists avai

Food Control

journal homepage: www.elsevier .com/locate/ foodcont

Short topical review

Mycotoxin management in the European cereal trading sector

David Siegel a,*, Teresa Babuscio b

aBAM Federal Institute for Materials Research and Testing, Richard-Willstätter-Str. 11, 12489 Berlin, GermanybComité du Commerce des céréales, aliments du bétail, oléagineux, huile d’olive, huiles et graisses et agrofournitures (COCERAL), Rue du Trône 98, 1050 Brussels, Belgium

a r t i c l e i n f o

Article history:Received 31 December 2010Received in revised form14 February 2011Accepted 22 February 2011

Keywords:MycotoxinsCerealsTradingAnalysisFood safety

Abbreviations: CEN, European Committee for Standmaximum level; DON, Deoxynivalenol; ELISA, Enzassay; GAFTA, The Grain & Feed Trade Association; GCHigh-performance liquid chromatography; IAC, ImMaximum level; MST, Member State; MS/MS, TandeOchratoxin A; RASFF, Rapid Alert System for Food an* Corresponding author. Tel.: þ49 30 8104 5889; fa

E-mail addresses: ds@dsscience.de (D. Siegel), te(T. Babuscio).

0956-7135/$ e see front matter � 2011 Elsevier Ltd.doi:10.1016/j.foodcont.2011.02.022

a b s t r a c t

On the world scale, the European Union has established the most comprehensive regulations formycotoxins in food and feed. These regulations, which are inter alia expressed in the form of maximumlevels, largely affect cereal traders. To ensure the safety of their products and compliance with EUlegislation traders are required to quantify the mycotoxin levels in their lots. However, while theanalytical approaches of research and enforcement are well known and frequently reviewed in thescientific and legal literature, little detailed information is available on the mycotoxin managementconcepts of trade. The present article is intended to close this gap. On the basis of the results of twosurveys conducted amongst European cereal traders in the years 2007 and 2009, three key issues incommercial mycotoxin management are outlined and discussed in the context of the current scientificliterature. These are: the issue of sampling, the availability and performance of suitable analyticalmethods as well as issues evolving from variations between regulatory and contractual maximum levels.

� 2011 Elsevier Ltd. All rights reserved.

Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11451.1. Mycotoxins and the European cereal trading sector . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11451.2. MLs in the EU . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11461.3. Consequences of ML violations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11471.4. Other legal provisions affecting cereal traders . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11471.5. Contractual provisions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1147

2. Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11483. Results and discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1148

3.1. Prevention . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11483.2. Sampling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11483.3. Chemical analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11493.4. Conflicts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1151

4. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1152Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1152References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1152

ardization; CML, Contractualyme-linked immunosorbent, Gas chromatography; HPLC,muno-affinity column; ML,m mass spectrometry; OTA,d Feed; ZON, Zearalenone.x: þ49 30 8104 1127.resa.babuscio@coceral.com

All rights reserved.

1. Introduction

1.1. Mycotoxins and the European cereal trading sector

Mycotoxins are toxic secondary metabolites produced bya range of fungi, including those genera which are colloquiallyreferred to as molds. Such fungi are able to infect a multitude ofhosts, as for instance cereals, whereas the accumulation of myco-toxins can take place in the field, during storage, processing or evenon final food and feed products.

D. Siegel, T. Babuscio / Food Control 22 (2011) 1145e11531146

The toxic effects triggered by the ingestion of mycotoxins varyfrom compound to compound; however, a chronic activity (e.g.carcinogenicity, teratogenicity or mutagenicity) is a commoncharacteristic in most cases (Bhat, Rai, & Karim, 2010; Wild & Gong,2010). Hence, in order to protect food consumers and livestock,mycotoxin levels in cereals need to be controlled. To do so, thecereal production sector has modified agricultural practices toavoid fungal growth (Kabak, Dobson, & Var, 2006). On the regula-tory side, legally binding, EU-wide maximum levels (MLs) formycotoxins in food were introduced by the European Commissionin 2001 and updated subsequently. For feed, a single ML for afla-toxin B1 and several guidance levels other mycotoxins were set(vide infra). While the food and feed sector accounts for the majorshare of the total cereal utilization in Europe (Table 1), further,industrial uses of cereals have gained importance in recent years.Such applications are, for instance, the production of bioethanol orbiopolymers from cereal based raw materials (Langeveld, Dixon, &Jaworski, 2010). In this context, it was pointed out that mycotoxinscan affect bioethanol production by inhibiting the employed yeasts(Boeira, Bryce, Stewart, & Flannigan, 2000; Klosowski, Mikulski,Grajewski, & Blajet-Kosicka, 2010). Further issues arise from thefact that the by-products of bioethanol production, which arefrequently used as feed, can havemycotoxin levels even higher thanthose of the raw material due to concentration effects (Biksey &Wu, 2009; Wu & Munkvold, 2008). For the sum of these reasons,cereal lots intended for industrial uses are often required to complywith quality criteria similar to those applicable to food and feed.

In any case, a conclusive statement on the quality of a cereal lotcan only be made when its mycotoxin content is known to a suffi-cient degree of accuracy. This inherently requires chemical analysis.However, speed, cost and accuracy of the available analyticalmethods vary widely and usually a compromise has to bemade, thenature of which depends on the context and the aim of the analysis.In this respect, the approaches taken by science and regulation arewell documented in the scientific and legal literature, while littledetailed information is available on the analytical approaches oftrade. This is despite the fact that the European cereal market, witha total volume of 324.8 million tons (agricultural year 2009/2010, cf.Table 1) is subjected to the most comprehensive mycotoxin regu-lations worldwide, involving several maximum levels in the lowparts per billion (ppb) range. Traders are required to operate in thisarea of tension both as buyers and sellers, not only consideringofficial regulations but also specific provisions of commercialcontracts. This leads to a high demand in practical and reliablemycotoxin analysis methods.

The commercial use of rapid mycotoxin test kits has recentlybeen discussed in detail (Alldrick, van Egmond, & Solfrizzo, 2009).However, although rapid test kits are of high importance in thecommercial sector, there are several other major issues which needto be addressed. We thus felt the need to explore the trade sector’s

Table 1Key figures of the European (EU-27) cereal market, compiled from COCERAL cerealbalance sheets. Cereals include soft and durum wheat, barley, maize, rye as well asoats, triticale and other minor cereals.

Agriculturalyear 2007e2008[million tons]

Agriculturalyear 2009e2010[million tons]

Total cereal market(production þ imports e exports)

309.0 (¼ 100%) 324.8 (¼ 100%)

Market sharesFeed 160.8 (52.0%) 165.1 (50.8%)Food/human consumption 89.1 (28.8%) 87.5 (26.9%)Seed production and other uses 57.2 (18.5%) 63.2 (19.5%)Bioethanol 1.9 (0.6%) 9.0 (2.8%)

approaches to mycotoxin management in a more comprehensivefashion. This was done on the basis of quantitative data gatheredthrough of two surveys amongst European cereal traders in theyears 2007 and 2009.

As a basis for the forthcoming discussion, a brief introduction tothe legal and contractual provisions relevant to cereal traders willbe given. Subsequently, by interpreting the survey results, threecurrent key issues in the field of commercial mycotoxin manage-ment will be outlined. By doing so we intend to document thecurrent situation in European cereal trading with respect tomycotoxin management and to raise awareness for the prevalent“analytical needs” of the sector.

1.2. MLs in the EU

The general principles of the EU legislation on contaminants infood were laid down in 1993 (The Council of the EuropeanCommunities, 1993). This legal act empowered the EuropeanCommission to take measures ensuring the protection of publichealth, including the introduction of MLs (Zmudzki & Wisniewska-Dmytrow, 2004). As a consequence, the Commission’s ScientificCommittee for Food (SCF) establishedMLs for aflatoxins, ochratoxinA and patulin in food by 2001 (European Commission, 2001). Thisinitial regulation replaced former national legislation. It wasupdated several times and substituted in 2006 by EU regulation1881/2006 (European Commission, 2006c) which was furtherupdated in 2007 and 2010 (European Commission, 2007, 2010b).Summa summarum, the EU has implemented the most compre-hensive regulations for food mycotoxins worldwide (van Egmond,2004; van Egmond, Schothorst, & Jonker, 2007). The establishedMLs (Table 2) are binding in all member states.

The major factors affecting an ML are the toxicity of therespective mycotoxin, its occurrence in food products and theintake of the concerned food products by the population (Wu,2004). Hence, the same mycotoxin may have different MLs indifferent cereals. Moreover, MLs depend on the state of processing.This may be illustrated using the example of fumonisin B1/2 MLs inmaize and derived products, which are 4000 mg/kg for unprocessedmaize, 1400e2000 mg/kg for different maize flours, 1000 mg/kg formaize based foods, 800 mg/kg for maize-based breakfast cereals andonly 200 mg/kg for maize based children/infant food (EuropeanCommission, 2006c, 2007).

For feed, the legal situation is somewhat different, with onlyaflatoxin B1 being regulated with binding MLs (range: 5e20 mg/kg)(European Commission, 2003; The European Parliament and theCouncil, 2002a). For the remaining toxins regulated in solidfoodsddeoxynivalenol (DON), zearalenone (ZON), ochratoxin A(OTA) and the fumonisins B1 and B2donly non-binding guidancevalues are set for feed (European Commission, 2006a). This is due tothe fact that, with the exception of the aflatoxins, contaminatedfeed does not directly or indirectly impact the human health, i.e.there is merely a negligible carryover to animal products (EuropeanCommission, 2006a).

In the case of lots intended for industrial purposes (e.g. bio-ethanol or biopolymer production) neither MLs nor guidance levelshave been established.

Mycotoxin MLs have a direct impact on all European food/feedbusiness operators and traders. The obligation for a proactivemycotoxin control on their part is legally rooted in EU Regulation178/2002, in which it is stated that “Food and feed business oper-ators at all stages of production, processing and distribution withinthe businesses under their control shall ensure that foods or feedssatisfy the requirements of food law which are relevant to theiractivities and shall verify that such requirements are met.” (TheEuropean Parliament and the Council, 2002b). Hence, routine

Table 2Current ranges of MLs and guidance levels for food and feed mycotoxins in the EU (August 2010) (European Commission, 2003, 2006a, 2006c, 2007, 2010b, 2010c;Weidenbörner, 2001).

Mycotoxin ML range: lowest e highest [mg/kg] Feed guidance level range:lowest e highest [mg/kg]

Main point of formation(non-exclusive)

Aflatoxins B1e2, G1e2

(sum of 4)0.10 (infant food) e 20 (feed for cattle, sheepand goats, except young animalsand dairy cows)

e In the field and during storage

Aflatoxin M1 0.025 (infant food) e 0.050 (milk) e Metabolic formation from aflatoxinB1 in dairy cows

Deoxynivalenol (DON) 200 (infant food) e 1750 (unprocessed maize) 900 (pig feed) e 12,000 (maize by-products) In the fieldFumonisins B1/2 (sum of 2) 200 (infant food) e 4000 (unprocessed maize) 5000 (feed for pigs, horses,

rabbits and pet animals) e 60,000(maize based feed)

In the field

Ochratoxin A (OTA) 0.50 (infant food) e 80 (liquorice extract) 50 (pig feed) e 250 (cereal based feed) During storage (particularlyin the case of cereals)

Patulin 10 (infant food) e 50 (fruit juices) e In the field (apples)Zearalenone (ZON) 20 (infant food) e 400 (refined maize oil) 100 (feed for piglets) e 3000

(maize by-products)In the field and during storage(e.g. on wet maize)

2 Before being adopted by CEN, methods are subject to an international, collab-orative validation study (Horwitz, 1995). The adoption process itself includes peer

D. Siegel, T. Babuscio / Food Control 22 (2011) 1145e1153 1147

mycotoxin analysis is essential for the purposes of both consumerprotection and legal compliance.

1.3. Consequences of ML violations

Although the economic consequences of an officially detectedmycotoxinML violation vary from case to case, a rough estimate canbe made for import controls done at an EU border. A figure for theaverage cost of an “EU border rejection” is given by Wu (2008). Itincludes testing and sampling, transportation, demurrage (storage,time, and labor costs), financial adjustments as well as reprocessingand amounts to €7300e11,000 (Wu, 2008). This figure, however,covers only administrative expenses. An ML violation can implyadditional costs, depending on the further fate of the concerned lot.In the special case of a lot violating the (low) MLs for food butcomplying with the (higher) guidance values for feed, it is feasibleto re-declare the lot as feed. The dilution of contaminated lots withuncontaminated material, on the other hand, is forbidden.1 Othermycotoxin reduction techniques applicable to contaminated foodlots, e.g. the use of adsorbants to remove/inactivate mycotoxins, arenot permitted. In the case of feed, additives can be employed forpurposes of detoxification, however, this use is limited to lots whichdo not exceed theMLs or guidance levels (The European Parliamentand the Council, 2002a, 2003). Hence, if a lot is contaminated tooheavily to be re-declared as feed, the EU officials can request itsdestruction, which results in a total loss for the carrier. There is noclear legal guideline for deciding whether a lot violating EU regu-lations is rejected, destroyed or accepted after re-declaration. Thisdecision is thus dependent on the administrative discretion of theEU officials conducting the control (cf. Article 18 of EC regulation882/2004 (The European Parliament and Council, 2004)).

The above considerations similarly apply to the intra-Europeanfood and feed market, which is monitored through official marketcontrols. However, differently to border controls, an intra-Europeanmycotoxins ML violationwill entail the tracing and identification ofthe source of contamination (The European Parliament and theCouncil, 2004). In 2009, a total of 6 border rejections and 8 intra-European alerts due to mycotoxins in cereals and bakery productswere reported to the Rapid Alert System of Food and Feed of the EU

1 This “non-dilution rule” does not apply to those feed mycotoxins subject toa guidance level rather than an ML. It furthermore only applies to lots which are“marketed for the first time”. This is, for instance, the case when a lot is delivered toan EU market from a silo. However, before testing and marketing, a farmer may fillthe silo with several batches of grains, which were not individually tested formycotoxins.

(cf. 443 border rejections and 39 alerts due to mycotoxins in nuts,nut products and seeds in the same timeframe) (EuropeanCommission, 2008).

1.4. Other legal provisions affecting cereal traders

A second aspect of EU mycotoxin legislation, indirectly affectingthe cereal processing industry and traders, concerns the techni-calities associated with ML surveillance and enforcement. The twoissues relevant in this context are sampling and analytical methodperformance. In both cases the legislator intends to define suitabletechniques and promote their application in all member states.There is, however, an important difference in the approaches taken:while in the case of sampling regulations provide concrete proto-cols, analytical methods are regulated merely in terms of theirperformance criteria, which include mycotoxin dependent accept-able ranges for a method’s recovery as well as maximum levels forits typical relative standard deviation. Actual methods are notrooted in EU regulations. However, the European Commissionrecommends the use of methods which have been established as“European Standards” (cf. Section 3.3).2

Regulations addressing the technical issues discussed aboveexist for food (European Commission, 2006b, 2010a) and feed(European Commission, 2009). They apply to all Europeanenforcement bodies, including local authorities, port healthauthorities and public analysts. While non-regulatory (i.e.commercial) analysts are not bound by the provisions, they maychoose to align their analytical methodologies with the methods ofthe enforcing bodies for the sake of consistency. This can help tobetter predict the outcome of an official control.

1.5. Contractual provisions

In addition to the official regulations, European cereal tradershave to consider provisions originating from contractual

reviews, consultations and an international voting (Gilbert & Anklam, 2002). TheEuropean Commission actively requests “European Standards” for mycotoxinanalysis as expressed in its mandate M 383 to CEN (European Commission, 2006d),in which it is stated that “the establishment of standardized methods of analysis isof utmost importance to guarantee a uniform application and control of theEuropean legislation in all Member States.” However, the CEN adoption has beencriticized for being too slow, causing the adopted methods to lack behind thetechnical state-of-the art (Gilbert, 1999).

Fig. 1. Survey question 1.

D. Siegel, T. Babuscio / Food Control 22 (2011) 1145e11531148

agreements with business partners. Business contracts in cerealtrading are mostly standardized and drafted by organizations likeThe Grain & Feed Trade Association (GAFTA, issuing standardizedcontracts for international trade) or several national organizationslike IncoGrain (France) and AGER (Italy).

With respect to mycotoxins, the contracts usually defineacceptable contractual maximum levels (CMLs). CMLs arecommonly oriented toward the official MLs, however, in some casesa client will request CMLs lower than the official ML. This is due toone major consideration: as outlined above, mycotoxin MLs arehigher for unprocessed raw materials than for processed foodproducts. The underlying reduction factor represents the extent ofmycotoxin reduction presumably achieved through the processingof raw materials. In the case example of fumonisins in maize, thelegally rooted reduction factor going from raw maize (ML:4000 mg/kg) to maize flour (particle size >500 mm, ML: 1400 mg/kg)is 35% (European Commission, 2007). Hence, the legislator expectsthat milled maize will contain merely 35% of the fumonisinspresent on the rawmaize. However, food processors can experiencedifferent, variable reduction factors. In the worst case scenario, theprocessed product will thus violate an ML although the rawmaterial complied with all relevant regulations. To avoid suchconflicts, the processing industries can (and will) request CMLsdistinctly below the official ML. Therefore, cereal traders may bebound by provisions even stricter than the ones established by theEuropean Commission.

CMLs are of particular importance also for lots intended forindustrial purposes, because mycotoxins can, for instance, interferewith yeasts used in the production of bioethanol (cf. Section 1.1).

In the case a lot is found to exceed a CML during a businesstransaction, a penalty or a cancellation of the transactionwill result.

2. Methods

Two survey rounds were done (round 1: 20th of June to 31st ofJuly 2007, round 2: 20th of August to 20th of September 2009), withthe questionnaire being revised and extended for the second round(2009). Traders from Austria, Belgium, Denmark, France, Germany,Italy, Poland, Portugal, Spain, Sweden, and the United Kingdomparticipated in the 2007 survey. In 2009, replies were obtainedfrom traders in Belgium, Denmark, France, Germany, Greece,Hungary, Ireland, Italy, the Netherlands, Portugal, Sweden, and theUnited Kingdom.

In the 2009 survey, each participant was asked to specify thevolume of cereals traded (comprising imports as well as localproduction) as well as the number of farmers represented by hisbusiness. The total volume of cereals traded (237,840,441 tons) andfarmers (1,965,350) in the involved EU member states (MSTs) in2009 were taken from COCERAL cereal balance sheets. Based onthese data, the received survey replies directly represented 21% ofthe cereal volume traded and 12% of the farmers active in theinvolved EU MSTs in 2009 (for the 2007 survey similar data werenot available).

The obtained replies were averaged for each EU MST, returningthe MST-dependant mean reply RMST [%]. The latter was assumed tobe representative for the whole cereal volume traded in therespective MST. The overall result for all MSTs (R [%]) was obtainedas the average of all RMST weighted by the individual MST’s share ofthe cereal volume traded by all participating MST’s:

R ¼ 1VAll

X

MST

VMSTRMST (1)

with R [%] ¼ overall result, VMST ¼ volume traded in MST,VAll ¼ volume traded in all MST’s, RMST ¼ mean reply for MST [%].

3. Results and discussion

3.1. Prevention

Once a lot is found to contain mycotoxins, it must not be diluted(cf. Section 1.3). Hence, prevention is the key strategy in avoidingmycotoxin related losses. There are several approaches successfullylowering the average contamination level of agriculturalcommodities pre- and post-harvest, however, it should be kept inmind that molds are ubiquitously occurring organisms, which cannever be fully eliminated under field conditions (Bhat et al., 2010).Typical pre-harvest prevention techniques (directed against fieldmycotoxins) include the use of mold resistant crop varieties, agri-cultural practices like field management as well as the use ofchemical and/or biological fungicides. Post-harvest methods(directed against storage mycotoxins) are usually focused on theimprovement of storage conditions, which can be achieved byeffective drying, temperature control and/or fungicide treatment(Kabak et al., 2006; Magan, Aldred, Mylona, & Lambert, 2010).

In the 2009 survey, 60% of the traders stated to advise theircontractual farmers on mycotoxin risk management (20% did notadvise, 20% did not reply to the question). Fig. 1 shows details onthe kinds of advice given. Subsequent to advising, 80% of the traderssaw a reduction of the mycotoxin problem (18% saw no changewhile 2% reported a worsening of the situation). These data stressthe high potential of prevention in managing mycotoxin relatedrisks.

3.2. Sampling

Once harvested, the crop needs to be subjected to chemicalanalysis in order to determine its contamination grade. This is fullyacknowledged amongst cereal traders: in the 2009 survey, 98% ofthe participants stated to test their lots for mycotoxins.

Since it is technically impossible to analyze an entire lot non-destructively, a representative sample is taken and analyzed. Thisseemingly simple sampling step, however, bears to highestpotential for errors in the whole analytical chain (Blanc, 2006; vanEgmond et al., 2007; Miraglia, De Santis, Minardi, Debegnach, &Brera, 2005; Whitaker, 2006). The main reason for this is theinhomogeneous distribution of mycotoxins in agriculturalcommodities and lots thereof, whereas the degree of inhomoge-neity depends on whether the mycotoxin was formed in the field(“field mycotoxins”, like DON, ZON and the fumonisins) or duringstorage (“storage mycotoxins”, like OTA), cf. Table 2. Froma sampling perspective, storage mycotoxins are more problematic,

Fig. 3. Typical variants of the analytical chain in commercial mycotoxin management.

D. Siegel, T. Babuscio / Food Control 22 (2011) 1145e1153 1149

because they are formed in “localized hot spots”, leading to a highlyinhomogeneous distribution (Casado, Parsons, Weightman, Magan,& Origgi, 2009). A concise illustration of this issue is given by Blanc,who stated that in a 10 kg lot, consisting of 20,000 peanuts, a single,highly contaminated kernel is sufficient to cause the whole lot toviolate the EU aflatoxin ML (Blanc, 2006). Hence, whether thissingle kernel is sampled or not will entirely determine the testresult, whichdin any casedwill not be representative for thewhole lot. To account for the pronounced influence of sampling onthe analytical results raised in the frame of official controls, the EUhas laid down detailed sampling plans for a range of raw materialsand foods in Commission Regulations 401/2006 and 178/2010(European Commission, 2006b, 2010a). However, while these planswere meant to harmonize official sampling regimes, they werecriticized for being unrealistic in terms of the needed workforce:e.g. two food inspectors would need half a working day for thesampling of just one container of nuts packed in sacks (van Egmondet al., 2007). Hence, the EU has issued a “Guidance document forthe sampling of cereals for mycotoxins” (European Commission,2010d) in which it is acknowledged that the provisions ofCommission Regulation 401/2006 are practically impossible toapply to e.g. cereal lots>500 tons stored inwarehouses or silos. Forproducts stored in small customer packages, the case is equallydifficult. The guidance document thus suggests alternate methodsof sampling. This is in accordance with Commission Regulation401/2006 which permits the use of alternate sampling methods incases of unacceptable commercial consequences or practicalunfeasibility of the official method (European Commission, 2006b).However, since the alternate methods are not rooted in the legalliterature, the official sampling procedures remain disharmonized.In this context, it is noteworthy that the European regulations donot define performance criteria for sampling plans (cf. Section 1.4).Hence, there is no requirement to establish the sampling contri-bution to the overall analytical uncertainty, which can remainlargely unclear. Consequently, while upon regulatory controls, theuncertainty contribution due to the chemical analysis step is knownand deducted from the test result in favor of the owner of the lot(cf. Section 3.3), this is not done for the sampling contribution. Therisk of wrongful lot rejections due to unrepresentative samples isthus borne by the traders in full. This risk (i.e. the analyticaluncertainty due to sampling) is not even known in many cases.

In order to reduce ambiguities due to sampling, traders areincreasingly adopting the existing regulatory sampling plans, thusaligning their methods with those used in enforcement (Fig. 2).However, individual sampling methods still play a major role. This

Fig. 2. Survey question 2.

is due to several reasons: Firstly, official sampling plans are oftenmodified to suit the special requirements and possibilities of thelogistic chains in the commercial sector. It is, for instance, recom-mended by the European Commission to sample grains “in-flow”,e.g. upon their transfer from a silo to a transport vehicle (EuropeanCommission, 2010d). However, this recommendation is not yetlegally rooted. Secondly, sampling regimes other than the officialones can be required by a business contract. And, finally, tradersmoving types of lots not mentioned in Commission Regulation401/2006 have to rely on “own methods” by default (Fig. 3).

3.3. Chemical analysis

Sampling processing is followed by chemical analysis. Althoughthe latter part of the analytical chain is less prone to error comparedto the sampling step, there are several complicating issues whichoften require compromises. This is mainly due to the variety ofanalytes and matrices (i.e. the components of the sample otherthan the analyte) under consideration. Mycotoxins are organicmolecules with varying physicochemical properties. At the sametime, the physicochemical properties of the sample itself varywidely, e.g. between different agricultural commodities like oatsand wheat. Due to this multitude of variables, there is no “goldstandard” method, which could cover all analytes and all matriceswith an equally sufficient performance (Köppen et al., 2010).

A second important issue concerns the conflict speed vs. accu-racy. While it is generally acknowledged that laboratory based,stationary methods produce the most accurate results (Anklam,Stroka, & Boenke, 2002; Köppen et al., 2010; Zheng, Richard, &Binder, 2006), they are also slowest and most expensive. In manysituations, however, traders are required to rapidly analyzea sample “on-site”, e.g. when a lot is received from a contractualpartner at a silo or warehouse (Fig. 4). Here, decisions of acceptanceusually need to be taken in “minutes rather than hours” (Alldricket al., 2009).

Fig. 4. Survey question 3.

Fig. 5. Survey question 4.

Fig. 7. Survey question 6.

D. Siegel, T. Babuscio / Food Control 22 (2011) 1145e11531150

The vast majority of portable methods used for mycotoxinanalysis are based on enzyme-linked immunosorbent assays (ELI-SAs) (Köppen et al., 2010; Krska & Molinelli, 2009; Zheng et al.,2006). Other, more recent rapid methods are based on fluores-cence polarization, for instance. In the case of ELISAs, qualitative,semi-quantitative and quantitative kits exist, whereas the speed ofanalysis usually decreases in the order given. In any case, a majoradvantage of the ELISA technique is its portability and ease of usecompared to stationary, chromatography based systems. Ready-to-use ELISA kits for mycotoxin analysis are offered by severalcompanies. However, a major disadvantage is the possibility of falsepositive results due to cross-reactions or false negative results dueto the inhibition of the ELISA antibodies by matrix components(Anklam et al., 2002). Studies on the risk associated with falsepositive or false negative results originating from rapid test kitshave been discussed in detail (Alldrick et al., 2009). It is importantto note that false results can occur not only upon using test kits ofdifferent manufacturers on the same sample, but also upon usingthe same test kit on different samples of the same agriculturalcommodity (Alldrick et al., 2009).

Laboratory-based mycotoxin analysis is almost exclusively doneusing HPLC coupled to various detectors like ultraviolet or fluo-rescence units as well as tandem-mass spectrometry (MS/MS)(Köppen et al., 2010; Turner, Subrahmanyam, & Piletsky, 2009). Asmost mycotoxins are non-volatile when not derivatized, gas-chro-matography (GC) is barely used. Contrary to ELISAs, HPLC based

Fig. 6. Survey question 5.

methods do not have to rely exclusively on biological materials(enzymes, antibodies) and hence produce more accurate results.Following on a cleanup by antibody-based immunoaffinity columns(IAC’s) an HPLC separation can largely eliminate false positiveresults due to antibody cross-reactivity, thus providing a maximumin selectivity. This advantage is reflected by the fact that themycotoxin methods, which have been adopted by the EuropeanCommittee for Standardization (CEN) as “European Standards” todate, are altogether based on HPLC, mostly in combination withIAC’s.3

As can be inferred from Figs. 5 to 7, the majority of traders avoidthe conflict “speed vs. accuracy” by relying both on rapid and HPLCbased methods. Although some alternatives have recently becomeavailable (Krska & Molinelli, 2009), ELISAs is still the rapid methodof choice. While ELISA kits are mainly used in-house due to theirlow cost and ease-of-use (Fig. 6), HPLC analyses are frequentlyoutsourced to external laboratories (Figs. 5 and 7). By comparingthe data for 2007 and 2009 it can be seen that this allocationbecame even more pronounced during the period under consid-eration. The major reasons for this are the steadily decreasing costof ELISA kits for mycotoxins and the increasing number ofaccredited, external laboratories offering routine mycotoxinanalysis.

The commissioning of ISO/IEC 17025 accredited laboratories isof particular importance to cereal traders (Fig. 8), as the ISO/IEC17025 standard inter alia ensures the correct reporting of theanalytical results obtained. In practice, incomplete reporting oftencauses ambiguity about a sample’s compliance with a given ML orCML. This is for instance the case when the commissioned

Fig. 8. Survey question 7.

Fig. 9. Survey question 8.Fig. 11. Survey question 10.

D. Siegel, T. Babuscio / Food Control 22 (2011) 1145e1153 1151

laboratory does not disclose whether the reported quantitativeresult was corrected for recovery or not. Further, the uncertainty ofthe analytical result needs to be established and communicated tothe client. Upon an official control, the expanded uncertainty(expansion factor 2, corresponding to a confidence level of 95%) ofthe test result must be deducted from the recovery corrected testresult itself in favor of the owner of the concerned lot (EuropeanCommission, 2006b). Hence, only if both recovery and expandeduncertainty are known, a sound decisionwith respect to ML or CMLcompliance can be made. Both the ISO/IEC 17025 standard as wellas Commission Regulation 401/2006 include provisions to ensurethat the respective information, which needs to be obtained bymeans of method validation, is forwarded to the client.

Concerning the analytes tested for, the data shown in Fig. 9reasonably reproduce the importance of the different mycotoxinsin European cereals. DON, ZON and OTA are the prevalent analytes,as they are known to occur on all major cereals relevant to trade,i.e. barley, maize, millet, oats, wheat, rice, rye, sorghum etc.(Weidenbörner, 2001). The lower analysis frequency for the afla-toxins and the fumonisins is due to the fact that these compoundsdo not play a major role on certain cereals like barley or wheat(Weidenbörner, 2001). In the special case of the T2/HT2 toxins,European regulations are not yet in place. Hence, the availability ofanalytical methods and rapid test kits is limited. This mightcontribute to a lower analysis frequency.

3.4. Conflicts

Mycotoxin related conflicts in cereal trading arise when officialor contractual controls indicate the violation of an ML or CML.

Fig. 10. Survey question 9.

Figs. 10e12 show that conflicts occur more frequently subsequentto official controls than to contractual controls. In any case,considering that 98% of the traders test their lots for mycotoxins(cf. Section 3.2), the number of yet occurring conflicts is surprising.As for the official controls, a possible explanation might be that theanalytical chains (Fig. 3) of trade and the enforcing bodies are notharmonized, thus producing varying results. This problem isaccentuated by the fact that there is a lack of harmonized samplingplans for cereals (cf. Section 3.2). Secondly, it can be noted that themycotoxin analysis methods adopted by CEN to date cover by farnot all of the mycotoxinematrix combinations for which MLs are ineffect (European Committee for Standardization, 2009). Also,portable techniques are not yet part of the CEN mycotoxin methodportfolio. These shortcomings make it difficult for any food or feedbusiness operator to harmonize or align their analytical chains withthose used by the enforcing bodies.

In the case of testing done by trading partners, the analyticalchain to be used both by buyer and seller is often very similar.Conflicts due to customer complaints are thus less frequent (Figs. 11and 12). Even so, there appears to be a residual risk which cannot beavoided byaligning analytical chains. This residual risk ismost likelydue to performance issues of the involved sampling plans oranalytical methods, leading to randomly altered test results.3

A conclusion, which can be drawn from all these aspects is, thatthere is a pronounced need for validating the analytical chains usedin commercial mycotoxin analysis. Validation is an essential tool forunderstanding a procedure’s performance and its influence on thetest result. While validation itself does not improve an analyticalprocedure, it is the only way to make a conclusive statement onwhether or not it is actually “fit for purpose”. Consequently, theultimate goal of validation is to characterize and lower both thebuyer’s and the seller’s risk in transactions relying on chemicalanalysis for quality control (Hubert et al., 2004). While the aligningof the analytical chains used by buyers and sellers (or traders andenforcing bodies) is an easily feasible way to produce moreconsistent results, it can not replace validation (even if the samemethods are used by both parties, the results can vary due toimprecision or lacking robustness of the method; also, identical,but wrong results can be obtained).

Unfortunately, there is a pronounced shortage of validatedsampling plans and analytical methods available to traders orcontractual laboratories. This introduces some arbitrariness to theanalytical results gathered for trade related decisions.

3 To avoid losses due to unpredictable test results, traders may effect an insur-ance policy. In 2009, 25% of the traders stated to have been covered by insurance ina mycotoxin related conflict.

Fig. 12. Survey question 11.

D. Siegel, T. Babuscio / Food Control 22 (2011) 1145e11531152

4. Conclusions

In this article, we have outlined mycotoxin related legal andtechnical challenges presented to European cereal traders on thebasis of a literature review and quantitative survey data. Inconclusion, we believe that the following three issues can beconsidered most substantial from a trade perspective:

(i) Trade has to consider both MLs and CMLs, whereas thedifference between the two arises from the concern about themycotoxin reduction factors practically achievable by cerealprocessing. This similarly applies to food, feed and industrialapplications. As CMLs are usually lower than official MLs,additional strain is put on cereal traders.

(ii) Sampling is the major source of uncertainty in mycotoxinanalysis (cf. Section 3.2). Therefore, harmonized, legallyaccepted sampling schemes for those types of commoditiesand lots relevant in cereal trading are required. As a next step,it would be highly desirable to quantify the degree of uncer-tainty introduced through those sampling schemes in thesense of a sampling plan validation. Such quantitative data aresurprisingly limited. This applies both to the legal and thescientific literature.

(iii) Although the European legislation on mycotoxins is harmo-nized since 2001 and although a range of performance criteriafor mycotoxin analysis methods have been legally defined,there still is a pronounced need for validated analyticalmethods. This need is particularly strong in the field of rapidmethods capable of supporting on-site decisions of lotacceptance. Due to the significant risk of false positive or falsenegative results obtained through the currently available,antibody-based rapid test kits, trade is often forced to rely onlengthy and costly confirmatory analyses by external labora-tories (Figs. 3 and 5). Improved rapid test kits, which areadequately characterized by means of comprehensive valida-tion, could largely reduce this burden. It should, however, beacknowledged that test kits need to be applied correctly, thatis only in the frame of their technical specifications and only tothe matrices they were validated for. In this respect, trainedpersonnel and an appropriate working environment arefurther fundamental prerequisites (Alldrick et al., 2009).

The issue of food and feed mycotoxins is, after all, by far notresolved and will continuously challenge regulators, scientists andfood business operators. In view of this, the present article isintended to contribute to a better understanding between theinvolved stakeholders. We believe that it should be a common

effort to reduce the many sources of uncertainty in modernmycotoxin analysis.

Acknowledgments

The authors would like to acknowledge Mr. Jean-Michel Aspar(chair of the COCERAL food and feed safety section), Mr. SebastienPicardat (Chair of the COCERAL working group on mycotoxins) andall other members of the COCERAL working group on mycotoxinsfor valuable comments and discussions. The authors are further-more grateful to Mrs. Gloria Gabellini (COCERAL) and to Mr. PaulRooke (COCERAL UK member) for the compilation of data used inthis article.

References

Alldrick, A. J., van Egmond, H. P., & Solfrizzo, M. (2009). Commercial use of rapidmycotoxin test kits: significance and potential harmonisation issues. WorldMycotoxin Journal, 2(2), 215e220.

Anklam, E., Stroka, J., & Boenke, A. (2002). Acceptance of analytical methods for imple-mentation of EU legislation with a focus onmycotoxins. Food Control, 13, 173e183.

Bhat, R., Rai, R. V., & Karim, A. A. (2010). Mycotoxins in food and feed: present statusand future concerns. Comprehensive Reviews in Food Science and Food Safety,9(1), 57e81.

Biksey, T. M., & Wu, F. (2009). Biofuels: by-products. Science, 326(5958), 1344e1345.Blanc, M. (2006). Sampling: the weak link in the sanitary quality control system of

agricultural products. Molecular Nutrition & Food Research, 50(6), 473e479.Boeira, L. S., Bryce, J. H., Stewart, G. G., & Flannigan, B. (2000). The effect of combi-

nations of fusariummycotoxins (deoxynivalenol, zearalenoneand fumonisinB-1)on growth of brewing yeasts. Journal of Applied Microbiology, 88(3), 388e403.

Casado, M. R., Parsons, D. J., Weightman, R. M., Magan, N., & Origgi, S. (2009).Geostatistical analysis of the spatial distribution of mycotoxin concentration inbulk cereals. Food Additives and Contaminants Part A-Chemistry Analysis ControlExposure & Risk Assessment, 26(6), 867e873.

van Egmond, H. P. (2004). Natural toxins: risks, regulations and the analyticalsituation in Europe. Analytical and Bioanalytical Chemistry, 378(5), 1152e1160.

van Egmond, H. P., Schothorst, R. C., & Jonker, M. A. (2007). Regulations relating tomycotoxins in food. Analytical and Bioanalytical Chemistry, 389(1), 147e157.

European Commission. (2001). Commission regulation (EC) 466/2001. OfficialJournal of the European Union, L 77, 1e13.

European Commission. (2003). Commission directive 2003/100/EC. Official Journalof the European Union, L 285, 33.

European Commission. (2006a). Commission recommendation (EC) 2006/576.Official Journal of the European Union, L 229, 7.

European Commission. (2006b). Commission Regulation (EC) 401/2006. OfficialJournal of the European Union, L 70, 12.

European Commission. (2006c). Commission regulation (EC) 1881/2006. OfficialJournal of the European Union, L 364, 5e24.

European Commission. (2006d). Mandate for standardisation addressed to CEN in thefield of methods of analysis for mycotoxins in food (M 383).

European Commission. (2007). Commission Regulation (EC) 1126/2007. OfficialJournal of the European Union, L 255, 14e17.

European Commission. (2008). Rapid alert system for food and feed (RASFF).https://webgate.ec.europa.eu/rasff-window/portal/ Access date: 19.11.2010.

European Commission. (2009). Commission regulation (EC) 152/2009. OfficialJournal of the European Union, L 54, 1.

European Commission. (2010a). Commission regulation (EC) 178/2010. OfficialJournal of the European Union, L 52, 32.

European Commission. (2010b). Commission regulation (EU) 105/2010. OfficialJournal of the European Union, L 35, 7e8.

European Commission. (2010c). Commission regulation (EU) No 165/2010. OfficialJournal of the European Union, L50, 8.

European Commission. (2010d). Guidance document for the sampling of cereals formycotoxins e final version. http://ec.europa.eu/food/food/chemicalsafety/contaminants/guidance-sampling-final.pdf Access date: 05.11.2010.

European Committee for Standardization. (2009a). CEN database of European stan-dards. Brussels.

Gilbert, J. (1999). Overview of mycotoxin methods, present status and future needs.Natural Toxins, 7(6), 347e352.

Gilbert, J., & Anklam, E. (2002). Validation of analytical methods for determiningmycotoxins in foodstuffs. TrAC, Trends in Analytical Chemistry, 21(6e7), 468e486.

Horwitz, W. (1995). Protocol for the design, conduct and interpretation of method-performance studies. Pure and Applied Chemistry, 67(2), 331e343.

Hubert, P., Nguyen-Huu, J. J., Boulanger, B., Chapuzet, E., Chiap, P., Cohen, N., et al.(2004). Harmonization of strategies for the validation of quantitative analyticalprocedures e a SFSTP proposal e part I. Journal of Pharmaceutical andBiomedical Analysis, 36(3), 579e586.

Kabak, B., Dobson, A. D. W., & Var, I. (2006). Strategies to prevent mycotoxincontamination of food and animal feed: a review. Critical Reviews in FoodScience and Nutrition, 46(8), 593e619.

D. Siegel, T. Babuscio / Food Control 22 (2011) 1145e1153 1153

Klosowski, G., Mikulski, D., Grajewski, J., & Blajet-Kosicka, A. (2010). The influence ofraw material contamination with mycotoxins on alcoholic fermentation indi-cators. Bioresource Technology, 101(9), 3147e3152.

Köppen, R., Koch, M., Siegel, D., Merkel, S., Maul, R., & Nehls, I. (2010). Determi-nation of mycotoxins in foods: current state of analytical methods and limita-tions. Applied Microbiology and Biotechnology, 86(6), 1595e1612.

Krska, R., & Molinelli, A. (2009). Rapid test strips for analysis of mycotoxins in foodand feed. Analytical and Bioanalytical Chemistry, 393(1), 67e71.

Langeveld, J. W. A., Dixon, J., & Jaworski, J. F. (2010). Development perspectives ofthe biobased economy: a review. Crop Science, 50(2), S142eS151.

Magan, N., Aldred, D., Mylona, K., & Lambert, R. J. W. (2010). Limiting mycotoxins instored wheat. Food Additives and Contaminants Part A-Chemistry Analysis ControlExposure & Risk Assessment, 27(5), 644e650.

Miraglia, M., De Santis, B., Minardi, V., Debegnach, F., & Brera, C. (2005). The role ofsampling in mycotoxin contamination: an holistic view. Food Additives andContaminants, 22, 31e36.

The Council of the European Communities. (1993). Council regulation (EEC) 315/93.Official Journal of the European Union, L 37, 1.

The European Parliament and Council. (2004). Regulation (EC) 882/2004. OfficialJournal of the European Union, L191, 1.

The European Parliament and the Council. (2002a). Directive (EC) 2002/32. OfficialJournal of the European Union, L 140, 10.

The European Parliament and the Council. (2002b). Regulation (EC) 178/2002.Official Journal of the European Communities, L 31, 1e24.

The European Parliament and the Council. (2003). Regulation (EC) No 1831/2003.Official Journal of the European Union, L 268, 29.

The European Parliament and the Council. (2004). Regulation (EC) 882/2004. Offi-cial Journal of the European Union, L 165, 1.

Turner, N. W., Subrahmanyam, S., & Piletsky, S. A. (2009). Analytical methods fordetermination of mycotoxins: a review. Analytica Chimica Acta, 632(2),168e180.

Weidenbörner, M. (2001). Encyclopedia of food mycotoxins (1st ed.). Berlin: Springer.Whitaker, T. B. (2006). Sampling foods for mycotoxins. Food Additives & Contami-

nants: Part A: Chemistry, Analysis, Control, Exposure & Risk Assessment, 23(1),50e61.

Wild, C. P., & Gong, Y. Y. (2010). Mycotoxins and human disease: a largely ignoredglobal health issue. Carcinogenesis, 31(1), 71e82.

Wu, F. (2004). Mycotoxin risk assessment for the purpose of setting internationalregulatory standards. Environmental Science & Technology, 38(15), 4049e4055.

Wu, F. (2008). A tale of two commodities: how EU mycotoxin regulations haveaffected US tree nut industries. World Mycotoxin Journal, 1(1), 95e102.

Wu, F., & Munkvold, G. P. (2008). Mycotoxins in ethanol co-products: modelingeconomic impacts on the livestock industry and management strategies. Journalof Agricultural and Food Chemistry, 56(11), 3900e3911.

Zheng, M. Z., Richard, J. L., & Binder, J. (2006). A review of rapid methods for theanalysis of mycotoxins. Mycopathologia, 161(5), 261e273.

Zmudzki, J., & Wisniewska-Dmytrow, H. (2004). Limits and regulations for myco-toxins in food and feed. Polish Journal of Veterinary Sciences, 7(3), 211e216.