The aim of this study was to develop a multicomponent analytical method for the determination of deoxynivalenol (DON), ochratoxin A (OTA) and zearalenone (ZEN), nivalenol (NIV), 3-acetyl-DON (3-acDON), 15-acetyl-DON (15-acDON), zearalenol (ZOL) and citrinin (CIT) in wheat. It also aimed to survey the presence and amounts of DON, OTA and ZEN in Belgian conventionally and organically produced wheat grain and in wholemeal wheat flours. After solvent extraction, an anion-exchange column (SAX) was used to fix the acidic mycotoxins (OTA, CIT), whilst the neutral mycotoxins flowing through the SAX column were further purified by filtration on a MycoSep cartridge. OTA and CIT were then analysed by high-performance liquid chromatography (HPLC) using an isocratic flow and fluorescence detection, while the neutral mycotoxins were separated by a linear gradient and detected by double-mode (ultraviolet light fluorescence) detection. The average DON, ZEN and OTA recovery rates from spiked blank wheat flour were 92, 83 and 73% (RSDR = 12, 10 and 9%), respectively. Moreover, this method offered the respective detection limits of 50, 1.5 and 0.05 microg kg-1 and good agreement with reference methods and inter-laboratory comparison exercises. Organic and conventional wheat samples harvested in 2002 and 2003 in Belgium were analysed for DON, OTA and ZEN, while wholemeal wheat flour samples were taken from Belgian retail shops and analysed for OTA and DON. Conventional wheat tended to be more frequently contaminated with DON and ZEN than organic samples, the difference being more significant for ZEN in samples harvested in 2002. The mean OTA, DON and ZEA concentrations were 0.067, 675 and 75 microg kg-1 in conventional samples against 0.063, 285 and 19 microg kg-1 in organically produced wheat in 2002, respectively. Wheat samples collected in 2003 were less affected by DON and ZEN than the 2002 harvest. Organic wholemeal wheat flours were more frequently contaminated by OTA than conventional samples (p < 0.10). The opposite pattern was shown for DON, organic samples being more frequently contaminated than conventional flours (p < 0.10).

Beer was chosen as a cereal-derived and homogeneous product for a comparison of organic and conventional production methods in terms of mycotoxin contamination levels. Ochratoxin A (OTA, a storage mycotoxin) and deoxynivalenol (DON, a field mycotoxin) were assessed by HPLC in organically and conventionally produced beers sold in Belgium. Immunoaffinity column (OchraTest and DONPrep) purification was used prior to HPLC analysis. For in-house validation, recovery experiments, carried out with the spiked beers in the ranges of 50-200 ng OTA l-1 and 20-100 microg DON l-1, led to the overall averages of 91% (RSD = 10%, n = 9) and 93% (RSD = 5%, n = 27), respectively. Organic beers collected during 2003-2004 were more frequently OTA-contaminated (95%, n = 40) than their conventional counterparts (50%, n = 40). Conventional beers were OTA-contaminated at a mean concentration of 25 ng l-1 (range: 19-198 ng l-1), while organic beers contained a mean level of 182 ng l-1 (range: 18-1134 ng l-1). High OTA contamination above the limit of 200 ng l-1 (up to 1134 ng l-1) occasionally occurred in organically produced beers. A complementary survey performed with the same brands in 2005 did not confirm this accidental presence of excessive OTA loads (range: 3-67 ng l-1 for 10 conventional beers and 19-158 ng l-1 for 10 organic beers). Establishing a maximum of 3 microg OTA kg-1 in malt, the application of the regulation EC No. 466/2001 (entered in force before the last sampling) may be related to the observed improvement. The overall incidence of DON was 67 and 80% in conventional and organic beers, respectively. DON concentrations ranged from 2 to 22 microg DON l-1 (mean = 6 microg DON l-1) in conventional beers, while organic beers ranged from 2 to 14 microg DON l-1 (mean=4 microg DON l-1). Thus, DON in beers does not appear to be a major matter of concern. From the statistical tests, it was concluded that the variation between different batches was significant (P < 0.0001), in contrast to that observed between different brands, showing a lack of homogeneity in the raw materials. This occurs either in organically or in conventionally produced materials. Considering these results, an optimized frequency of controls according to European Regulations EC No 466/2001 and EC No 856/2005 should be recommended to reject the irregular batches.

A total of 205 cornflake samples collected in Belgian retail stores during 2003-2004 were surveyed for the natural occurrence of fumonisin B1 (FB1), B2 (FB2), and B3 (FB3). These cornflake samples, originating from conventional as well as from organic production, were analyzed using an intralaboratory-validated LC-MS/MS method. Additionally, 90 cornflake samples were subjected to rapid screening using a flow-through enzyme immunoassay method to demonstrate the practicability of a screening test coupled to a validated confirmatory LC-MS/MS method for the management of food safety risks. FB(1) concentrations ranged from not detected (nd) [LOD (FB1) = 20 microg/kg] to 464 microg/kg with mean and median concentrations of respectively 104 +/- 113 and 54 microg/kg. For FB2 and FB3, the concentration ranges varied respectively from nd [LOD (FB2) = 7.5 microg/kg] to 43 microg/kg and from nd [LOD (FB3) = 12.5 microg/kg] to 90 microg/kg. Mean concentrations for FB2 and FB3 were respectively 12 +/- 8 and 21 +/- 15 microg/kg, while the median concentration was 11 microg/kg for FB2 and 19 microg/kg for FB3. From the statistical tests (chi2 and ANOVA model III), it could be concluded that the agricultural practice did not have any significant effect on the fumonisin concentrations but that the variation between different batches was significant (p < 0.0001).