The term mycotoxin designates a chemically diverse group of secondary fungal metabolites [1,2,3], mainly produced by species of the Aspergillus, Penicillium or Fusarium genus [4,5]. Depending on classification, 300–400 mycotoxins are known today [6]. Contamination may occur in the field as well as during (improper) storage and is largely dependent on environmental factors [7,8]. When ingested by humans or animals above a certain concentration, mycotoxins will cause a toxic response referred to as mycotoxicosis [1,4,5]. The symptoms elicited by mycotoxin consumption range from reduced animal productivity (reduced body weight gain, reduced fertility) and immune suppression [9], resulting in increased susceptibility to diseases and parasites to overt disease and death. Clinical symptoms of mycotoxin intoxication include diarrhoea, liver and kidney damage, pulmonary oedema, vomiting, haemorrhaging and tumours [1,7]. Under field conditions, mycotoxins usually occur in concentrations leading to reduced animal performance and/or immune suppression without causing any obvious clinical symptoms [5,8]. It is important to emphasise that it is very common for an array of mycotoxins to occur together at low concentrations. This is on the one hand due to the ability of various fungi to simultaneously produce a variety of mycotoxins [10,11] and on the other hand due to the fact that any given commodity is likely to be infected with different types of fungi. Moreover, compound feed is made up of a number of different commodities contributing to the final mycotoxin profile.

In the feed manufacturing process, only a limited number of mycotoxins are subject to regular testing and legal regulations/guidance [12,13], namely aflatoxins (AF), deoxynivalenol (DON), zearalenone (ZEN), fumonisins (FB) and ochratoxin A (OTA). Consequently, this review will focus on these contaminants. In addition, the case of T-2/HT-2 toxin will be addressed. First, a brief introduction about each of the aforementioned toxins including associated adverse effects will be provided. Subsequently, an overview of their occurrence and co-occurrence, with special focus on the EU, will be given.

In the European Union, maximum levels for aflatoxin B₁ in feedstuff have been established by the European Parliament [53] and guidance values for DON, ZEN, FB and OTA were issued by the Commission [13]. Selected examples are given in Table 1. Legally binding regulations (as opposed to mere guidance values) have been issued for AFB₁ as this mycotoxin is particularly prone to carry-over in milk (AFM₁). The maximum level is designed to prevent the occurrence of AFM₁ at levels considered harmful to human health.

A number of studies on the occurrence of mycotoxins in European feedstuffs have been published. Examples are given in Table 2. As this review aims at providing an update on mycotoxin contamination in feed, only publications dating from 2004 or later were included in this table. The data clearly show that mycotoxins are ubiquitously present in feed material throughout Europe and that maximum contamination levels exceeding the EU maximum levels or guidance values are likely to occur.

Miraglia et al. [54] emphasised that occurrence patterns of mycotoxins in Europe are expected to change as a consequence of rising average temperatures. Southern Europe is used as an example to support this statement. The authors report that while the importance of DON is about to decrease, A. flavus infection and aflatoxin contamination, previously uncommon in Europe, will become increasingly important. In fact, in 2003 a hot and dry growing season led to severe A. flavus infection of maize in northern Italy [55]. A survey on 110 samples showed an AFB₁ incidence of 75% with a mean contamination of 4.4 µg/kg. Using this maize as feedstuff for dairy cattle led to a widespread AFM₁ contamination in milk and several thousand tons of milk exceeding the EU legal limit of 0.05 µg/kg AFM₁ [12] had to be discarded. The study by Decastelli et al. [56] presents the results of a surveillance plan spanning the two years following this incident (Table 2). Indicators of this impending change were also found by Goertz et al. [57], who reported F. verticillioides, commonly associated with warmer and drier regions such as Italy or Spain, to be the predominant Fusarium species isolated from maize grown in Germany in 2006. As a result, FB contamination was detected in 34% of the samples (Table 2). According to the weather data provided in the report, July and September 2006 had been particularly warm and dry all over Germany.

It is however difficult to infer trends or recent developments regarding mycotoxin contamination in European feed from the data presented in Table 2. There are a number of reasons for this, not the least of which is the strong influence of the respective cropping season’s climate on the contamination level, causing a high year to year variation of the results, which underlines the importance of implementing regular surveillance programs. Also, applied analysis methods were diverse, including ELISA, TLC, GC and HPLC. Moreover sampling methods are often not described in detail, yet, sampling is considered the largest source of error in mycotoxin analysis [77].

The vast majority (>98%) of mycotoxin related entries concerning feed in the RASFF (Rapid Alert System for Food and Feed) database [78] report the detection of aflatoxins. This is not surprising, as aflatoxin B₁ is to date the only mycotoxin for which legal maximum levels in feedstuff have been established in the European Union. In general, mycotoxins only account for a rather small fraction of the feed related entries in the RASFF database [79], probably for the same reason.

Austrian feed and feed raw materials generally show good compliance with EU regulations and guidance levels. DON is detected most often (around 60% positives in cereal samples other than maize and 95% positives in maize samples). ZEN is the second most common mycotoxin, found in about 70% of unprocessed maize kernel samples. It is rarely detected in other cereal samples. Other mycotoxins such as FB and OTA are detected much less frequently, FB for example is only quantifiable in 10%–30% of each year’s maize samples [80,81,82]. One Austrian maize sample in 2009 and one in 2010 were found to exceed the guidance level for DON [80,81].

The vast majority of the feed samples analysed for mycotoxin contamination in 2010 within the scope of German official controls proved to be compliant with legal requirements. Seven oilseed samples were found to contain AFB₁ above the maximum limit. As 1810 samples were tested for AFB₁, this makes for only 0.4% of objectionable samples [83]. Two samples exceeded the guidance value for DON and three samples contained OTA levels above the respective guidance value [84].

Feed official control system in Greece reported a high compliance to EU regulation 1881/2006 in the majority of samples tested in 2010. For the detection of aflatoxin B₁, 71 feed and grain samples (wheat, maize, barley etc.) were tested and 5 were non-compliant to EU regulation 32/2002. All samples analysed for OTA, ZEN, and DON were compliant to the respective legislation and no excess levels of T-2 toxin were found either [85].

RIKILT Institute for Food Safety published a report analysing the results from Dutch official controls of feed from 2001–2009 [86]. Mycotoxin concentrations are generally well below the respective maximum level or guidance values although some individual samples exceeded admissible limits. Contamination levels in cereals were generally stable. High DON concentrations were found in liquid pig feed samples, 10% of which exceeded the guidance value. Oil seeds (sunflower, groundnut) and rice were reported to occasionally contain AFB₁ concentrations above the regulatory limit. It was also noted, that soy bean samples with high ZEN content often originated from Argentina.

In Sweden, all feed samples analysed under the national monitoring scheme in 2010 were compliant with EU regulations [87]. In fact, the Swedish board of agriculture did not detect any transgressions of mycotoxin maximum levels or guidance values since 2006 [88,89,90]. In early 2006, aflatoxin contaminated rice meal used in dairy cattle feed production had been identified as the cause of elevated AFM₁ levels in Swedish milk [91].

Considering the fact that mycotoxigenic fungi are usually capable of producing more than one mycotoxin [4,10,11] and that feed raw materials are commonly infected with various fungal species at a time [10], studying the occurrence of any given mycotoxin alone provides incomplete information about the risk associated with the respective feedstuff. Compound feed is particularly vulnerable to multiple contaminations as it typically contains a mixture of several raw materials [92]. In their meta-analysis, Grenier and Oswald [49] reviewed 112 publications on toxicological interactions of mycotoxins. They found that most of the studies reported synergistic or additive interactions regarding adverse effects on animal performance. When it comes to other parameters, especially biochemical ones, results were more variable, ranging from synergistic to antagonistic for the same toxin combination and target variable. Nevertheless the summarised findings underline the importance of studying mycotoxin co-occurrence.

Stoev et al. [93] arrived at a similar conclusion. They analysed a total of 50 feed samples from Bulgarian pig and poultry farms experiencing nephropathy problems. They found that, although all samples were contaminated with OTA at a mean concentration exceeding the EC guidance levels for pig and poultry feed (50 µg/kg and 100 µg/kg respectively) [13], OTA concentration alone was not high enough to explain the observed symptoms. It was inferred that the nephropathy problems were caused by the combined effects of OTA, FB₁ and penicillic acid. Furthermore, a hitherto unknown metabolite found in 92% of the samples, is suspected as an additional causative agent.

Both publications highlight the necessity of testing for an array of mycotoxins in order to accurately determine feed quality and potential risks. This section is aimed at providing an overview of mycotoxin co-occurrence found in feed and feed raw materials with special focus on European data. The findings mentioned in the following paragraphs are also summarised in Table 3 and Table 4. Table 4 contains data on mycotoxin co-occurrence in Europe while Table 3 provides additional information concerning the co-occurrence of AF and FB in maize from South America. Maize is particularly prone to infection by mycotoxigenic fungi, most notably Fusarium sp. [94]. Therefore it is often subject to studies on mycotoxin contamination and co-occurrence.

As for Europe, an UK study screening maize products intended for animal feed for 22 mycotoxins found that all 67 samples were co-contaminated with up to 12 different Fusarium mycotoxins [95]. FBs and DON occurred together in 75% of the samples and 15-AcDON, moniliformin, and ZEN were frequent co-contaminants. None of the samples contained detectable amounts of AF. Prior to this study, Scudamore et al. [96] conducted a survey on 330 samples of feed ingredients and found that maize was by far the most affected by co-contamination. In total, 60% of the maize samples tested positive for more than one mycotoxin, AF and FB being the most common combination (28%). Another publication reported the occurrence of 14 different Fusarium mycotoxins in 84 German maize samples sourced after the 2006 and 2007 harvest [57]. Although no information regarding co-occurrence is provided, the high individual incidences of several mycotoxins imply co-occurrence of DON (and its acetylated forms), ZEN, moniliformin, beauvericin, nivalenol, enniatin B, FBs, and/or HT-2 toxin. Likewise, co-occurrence of trichothecenes with ZEN and OTA in 760 Hungarian maize samples can be deduced from the individual toxin prevalences reported by Rafai et al. [97]. This study also comprised 921 samples of other feed raw materials such as wheat, barley, soybean and so on. Typically, DON was found to be the major contaminant, frequently co-occurring with ZEN.

Several groups investigated the occurrence of aflatoxins and fumonisins in Brazilian maize. The results are summarised in Table 3. Joint contamination with these toxins is particularly concerning as there is evidence that FB₁ synergistically promotes liver tumours initiated by AFB₁ [49]. Brazil is the third most important producer of maize in the world [98]. Hence, contamination levels of Brazilian maize are of international significance. Table 3 also lists a long-term study from Argentina, fourth most important producer of maize worldwide [98].

A survey on feed and feed ingredients (n = 416) sourced in Southern Europe from January 2005 to August 2009 established that 22% of the compound feed samples contained more than one mycotoxin [59]. It was furthermore reported that 23% of all samples from Spain and 32% of the Italian samples contained at least two mycotoxins. The samples were tested for AFs, ZEN, B-trichothecenes (DON and AcDON), A-trichothecenes (T-2 toxin and HT-2 toxin), FBs, and OTA. It was stated that Fusarium mycotoxins, namely type B-trichothecenes, ZEN, and FBs, were major contaminants and that their co-occurrence was frequently observed. Almeida et al. [62] analysed 277 samples of feed for fattening pigs marketed in Portugal for OTA, ZEN, and DON. In total, 10% of the samples were found to contain any two-toxin combination, ZEN and DON co-occurring most often. One sample was contaminated by all three toxins. Moreover, concentrations of FBs and ZEN were determined in 127 samples of sows feed, two of which (1.5%) were positive for both analytes. Another study analysing 82 samples of sows feed, wheat and maize sourced in different EU countries in 2008 for the presence of 23 different mycotoxins found that 75% of the samples contained more than one mycotoxin [58]. The majority of these samples contained DON, AcDON, and FBs in combination with other mycotoxins. Of 50 Slovakian poultry feed samples tested for trichothecenes and ZEN, 84% contained more than one mycotoxin. A four toxin combination of DON, ZEN, T-2, and HT-2 was detected most frequently (32%) [71]. A similar study on the occurrence of FB₁, FB_2, and moniliformin in poultry feed from Slovakia found that 25 (50%) of the samples contained all three mycotoxins [70]. The description of the samples in both publications strongly suggests, that the samples analysed in both studies were identical. Merging the results shows that all samples were contaminated with at least three mycotoxins and that 82% contained five or more. The most common combinations being DON, ZEN, T-2, HT-2, FB₁, and FB₂ and DON, ZEN, T-2, HT-2, FB₁, FB₂, and moniliformin, found in 14% and 12% of the samples respectively. Upon investigating the diets of dairy cattle in the Netherlands, Driehuis et al. [104] found that DON and ZEN were co-occurring in 44% of the diets, silage and compound feed being the major sources of exposure.

Other European studies have focused on a single commodity. For example, Ibáñez-Vea et al. [64] tested 123 Spanish barley samples from the 2007 and 2008 harvest for AFs, ZEN, and OTA. Detectable levels of AFB₁ were reported for all the samples and it was found to co-occur with OTA in 31% of the samples. Co-occurrence of AFB₁ and ZEN was 12% and AFB₁, ZEN, and OTA co-contamination was detected in 27% of the samples. The authors published a second study on the occurrence of type A and B trichothecenes in these samples [65]. DON was present in 95% of them and 43% contained three or more trichothecenes. Combining the results of both studies 96% of the samples were found to be contaminated with three or more mycotoxins [105]. The combinations AFB₁, OTA and DON and AFB₁, OTA, DON, and ZEN were most frequent, being observed in 29% and 26% of the samples respectively.

The frequent detection of mycotoxin co-occurrence even in studies screening for a limited number of analytes underlines the importance of multi-mycotoxin analysis methods. Berthiller et al. [6] published a comprehensive review on the evolution of such methods from HPLC-UV to state of the art LC-MS/MS. Of the reports summarised above, Goertz et al. [57], Monbaliu et al. [58] and Driehuis et al. [104] used LC-MS/MS methods for the simultaneous detection of various mycotoxins. A very powerful method to this end was developed by Sulyok et al. who extended their LC-MS/MS method published in 2006 [106] to detect 87 mycotoxins [2]. Upon further development the method was extended by another 19 mycotoxins [107]. Today, it allows for the simultaneous detection of 340 analytes, most of them being mycotoxins (Personal Communication, M. Sulyok).

Mycotoxins are ubiquitously present in European feed and feed raw materials. Although compliance with EU regulations is usually high, continuous monitoring is needed in order to avoid negative impacts on animal health and performance due to elevated contamination levels, which may occur for example after growing seasons characterised by weather conditions particularly favourable to fungal infection and growth. As the timing of the rainfall may be more important than the actual amount of precipitation [20,22], the development of predictive models for mycotoxin occurrence based on regional weather data would be a valuable tool to estimate the risk of contamination after a given growing season. For example, DONcast^®, a predictive tool for DON in wheat in Ontario (Canada), has been commercialised in 2002 [108]. A European version (DONcast^® Europe) is under validation (www.doncast.eu). In Italy, a risk assessment tool predicting FB production by F. verticillioides in maize is under development [109]. Although the weather is the most influential parameter regarding the extent of mycotoxin contamination (it was found to explain 48% of the variation in DON concentration observed in wheat and maize in a Canadian study [110]) other factors are important as well. These are for example the choice of variety, crop rotation (avoiding maize as a pre-crop for wheat), tillage (ploughing reduces inoculum from plant residues), or the planting date (earlier planting of maize is generally preferable). Jouany [94] provided a comprehensive review on this subject. Some studies also investigated whether organically farmed products were more prone to mycotoxin contamination resulting from the absence of pesticide use. Conflicting results have been obtained on this issue, suggesting that other factors may override the effect of chemical control of fungal infection on the prevention of mycotoxin formation [94]. By the same token, there is no indication that organically farmed animal products should be considered any more or less risky in terms of mycotoxin contamination than those produced by conventional farming [111].

It is also necessary to increase the farmer’s awareness to the issue of mycotoxins in feed. Piva et al. [55] reported that northern Italian farmers had been warned about elevated AFB₁-levels in maize after the 2003 growing season but that the warning was dismissed until the first detection of AFM₁ in milk. The occurrence of AFB₁ at such high levels in Europe also underlines the fact that climate change will entail a change in the mycotoxin distribution patterns observed today.

With regard to the frequent detection of mycotoxin co-occurrence, even if only a very limited number of mycotoxins are analysed, and the evidence of possible additive or synergistic interaction of co-occurring mycotoxins, guidelines or maximum levels should not only be set for each mycotoxin individually but also for particularly concerning combinations which would definitely require more data on the impact of different mycotoxin combinations on different animals species. Possible synergistic interactions with so-called emerging mycotoxins such as moniliformin, beauvericin or enniatins should not be neglected in this respect.

Although a lot of efforts to prevent mycotoxin formation have been undertaken, contaminations of those secondary fungal metabolites still occur. Therefore mycotoxin reduction strategies like the addition of mycotoxin deactivating products based on different strategies (adsorption, biotransformation, biodegradation, bioprotection) should be considered. Also, the development and application of multi-mycotoxin LC-MS/MS methods should be encouraged in order to get a more accurate picture of the extent of multi-mycotoxin contamination. These methods are especially valuable when it comes to detecting masked mycotoxins. Essentially, the term refers to conjugates of mycotoxins that typically go undetected when testing for the parent toxin. These conjugates may be produced by the fungus itself (3-AcDON, 15-AcDON) or formed as a part of the infected plant’s defence mechanism (DON-3-Glucoside, ZEN-4-Glucoside) [112]. Apart from exerting toxic effects themselves, there is evidence that some conjugates may be converted into the parent toxin during digestion, further adding to the toxicity of the feed [113,114].