A total of 230 samples of processed rice and its sub-products or derived products were analysed to establish the co-occurrence of several mycotoxins. Samples were analysed in the period 2007-2009 due to the outbreak of beriberi associated with the consumption of rice stored in inappropriate conditions in Brazil. According to data from the Ministry of Health, 323 cases of disease were registered in 2006, of which at least 47 cases resulted in death. The occurrence of total aflatoxin (AFT) (aflatoxin B(1) + B(2) + G(1) + G(2)), ochratoxin A (OTA), zearalenone (ZON), deoxynivalenol (DON), and citreoviridin (CTV) was 58.7%, 40.0%, 45.2%, 8.3% and 22.5%, respectively. From 166 rice samples analysed, 55% had levels <0.11 µg kg(-1) for AFT. For OTA and ZON, of 165 rice samples analysed, 28% and 29% were contaminated with levels from 0.20 to 0.24 µg kg(-1) and from 3.6 to 290.0 µg kg(-1), respectively. One sample (0.6%) was contaminated with 4872.0 µg kg(-1) of ZON. A total of 91% of rice samples (n = 165) did not contain detectable DON (<30.00 µg kg(-1)), although the highest level of contamination was found to be 244 µg kg(-1). From the total of 65 samples analysed, 94% had no detectable CTV (<0.9 µg kg(-1)), with a range from 0.9 to 31.1 µg kg(-1) in 6% of the samples. The highest levels of contamination were found in rice sub-products or derived products from the husk and rice bran. Co-occurrence was observed for AFT and ZON in 17.0%, AFT and OTA in 24.2%, AFT and CTV in 6.2%, OTA and CTV in 4.6%, and ZON and CTV in 3.1%. These fractions were also the major contributors for the co-occurrence. The results found show the necessity of monitoring rice production.

An immunoaffinity clean-up-based method for determining ochratoxin A (OTA) in green coffee aiming at one-dimensional thin layer chromatography (TLC) analysis was established. OTA was extracted with a mixture of methanol and aqueous sodium hydrogen carbonate solution, purified through an immunoaffinity column, separated on normal or reversed-phase (RP) TLC plates and detected and quantified by visual and densitometric analysis. The linear equation of the standard calibration curve by densitometric analysis gave R(2) > 0.999 (0.04-84 ng). The mean recovery (R) of OTA from spiked samples (1.8-109 microg kg(-1)) by densitometric and visual analyses were 98.4 and 103.8%, respectively. The relative standard deviations (RSD) for densitometric and visual analysis varied from 1.1 to 24.9% and from 0.0 to 18.8%, respectively. The RSD for naturally contaminated samples by densitometry (three levels of contamination, n = 3) varied from 11.1 to 18.1%. The correlation (R(2)) between high-performance liquid chromatography (HPLC) and densitometry, and between visual and densitometric analysis for spiked samples were > 0.99. The limit of detection (LOD) of the method was 0.5 microg kg(-1) for normal TLC. Toluene-ethyl acetate-88% formic acid (6:3:1 v/v/v) and acetonitrile-methanol-water-glacial acetic acid (35:35:29:10 v/v/v/v) were regarded as the suitable TLC solvents for eluting both standards and samples on normal and RP TLC plates, respectively. Toluene-acetic acid (99:1 v/v) was chosen as the spotting solvent among several others for giving the best sensitivity and resolution of OTA on TLC plates as well as the best recovery of OTA from standard and sample extract residues. Preliminary studies were carried out to investigate the reuse of the immunoaffinity column and the interference of caffeine in the OTA recovery.

A collaborative study was conducted to evaluate a method using immunoaffinity column cleanup with liquid chromatography (LC) for the determination of ochratoxin A (OTA) in green coffee at levels that could be included in possible future regulations of the European Union. The test portion was extracted with methanol–3% aqueous sodium hydrogen carbonate solution (50 + 50, v/v). The extract was filtered, and the filtrate was diluted with phosphate-buffered saline and applied to an immunoaffinity column containing antibodies specific for OTA. After washing, the toxin was eluted from the column with methanol and quantified by LC with fluorescence detection. Pairs of 4 homogeneous noncontaminated and naturally contaminated materials (mean levels of &lt;0.12, 2.44, 5.15, and 13.46 ng/g) and blank samples (&lt;0.12 ng/g) for spiking were sent to 20 participant laboratories from 8 countries. The materials were analyzed according to the method description and all difficulties encountered in the analysis were reported. Statistical analysis was carried out according to the Harmonized Protocol of the International Union of Pure and Applied Chemistry. The relative standard deviation for repeatability (RSDr) ranged from 7.42 to 20.94%, and the relative standard deviation for reproducibility (RSDR) ranged from 16.34 to 29.17%. The method showed acceptable within-laboratory and between-laboratories precision for green coffee materials, as evidenced by HorRat values of ≤0.85, at the studied range, for spiked and naturally contaminated materials. The mean recovery was 92.8% for green coffee material spiked with OTA at a level of 4.82 ng/g.

A study was conducted on the risk from aflatoxins associated with the kernels and shells of Brazil nuts. Samples were collected from processing plants in Amazonia, Brazil. A total of 54 test samples (40 kg) were taken from 13 in-shell Brazil nut lots ready for market. Each in-shell sample was shelled and the kernels and shells were sorted in five fractions: good kernels, rotten kernels, good shells with kernel residue, good shells without kernel residue, and rotten shells, and analysed for aflatoxins. The kernel:shell ratio mass (w/w) was 50.2/49.8%. The Brazil nut shell was found to be contaminated with aflatoxin. Rotten nuts were found to be a high-risk fraction for aflatoxin in in-shell Brazil nut lots. Rotten nuts contributed only 4.2% of the sample mass (kg), but contributed 76.6% of the total aflatoxin mass (µg) in the in-shell test sample. The highest correlations were found between the aflatoxin concentration in in-shell Brazil nuts samples and the aflatoxin concentration in all defective fractions (R(2)=0.97). The aflatoxin mass of all defective fractions (R(2)=0.90) as well as that of the rotten nut (R(2)=0.88) were also strongly correlated with the aflatoxin concentration of the in-shell test samples. Process factors of 0.17, 0.16 and 0.24 were respectively calculated to estimate the aflatoxin concentration in the good kernels (edible) and good nuts by measuring the aflatoxin concentration in the in-shell test sample and in all kernels, respectively.

No abstract