Determination of trichothecenes in cereals and cereal-based products by liquid chromatography-tandem mass spectrometry

A sensitive, accurate and precise method for the simultaneous determination of nivalenol (NIV), deoxynivalenol (DON), T-2 toxin (T-2) and HT-2 toxin (HT-2) in different food matrices, including wheat, maize, barley, cereal-based infant foods, snacks, biscuits and wafers, has been developed. The method, using liquid chromatography coupled with atmospheric pressure chemical ionization triple quadrupole mass spectrometry (LC-APCI-MS/MS), allowed unambiguous identification of the selected trichothecenes at low µg per kg levels in such complex food matrices. A clean-up procedure, based on reversed phase SPE Oasis® HLB columns, was used, allowing good recoveries for all studied trichothecenes. In particular, NIV recoveries significantly improved compared to those obtained by using Mycosep® #227 columns for clean-up of the extracts. Limits of detection in the various investigated matrices ranged 2.5-4.0 µg kg^-1 for NIV, 2.8-5.3 µg kg^-1 for DON, 0.4-1.7 µg kg^-1 for HT-2 and 0.4-1.0 µg kg^-1 for T-2. Mean recovery values, obtained from cereals and cereal products spiked with NIV, DON, HT-2 and T-2 toxins at levels from 10 to 1000 µg kg^-1, ranged from 72 to 110% with mean relative standard deviation lower than 10%. A systematic investigation of matrix effects in different cereals and cereal products was also carried out by statistically comparing the slopes of standard calibration curve with matrix-matched calibration curve for each of the four toxins and the eight matrices tested. For seven of the eight matrices tested, statistically significant matrix effects were observed, indicating that, for accurate quantitative analysis, matrix-matched calibration was necessary. The method was applied to the analysis of 57 samples of ground wheat originated from South Italy and nine cereal food samples collected from retail markets.     

Natural co-occurrence of deoxynivalenol and fumonisins B1 and B2 in Italian marketed foodstuffs

A survey was performed to obtain the frequency and levels of contamination by deoxynivalenol (DON) and fumonisins B₁ and B₂ (FB1, FB2) mycotoxins in Italian marketed foods. Of 202 samples investigated, including raw materials and processed cereal foods (bread, pasta, breakfast cereals, biscuits, baby and infant foods), 84% were contaminated with DON at levels from 0.007 to 0.930 μg g^-1 (median 0.065 μg g^-1); 26% contained FB1 ranging from 0.010 to 2.870 μg g^-1 (0.070 μg g^-1); 35% contained FB2 at 0.010-0.790 μg g^-1 (0.080 μg g^-1). The highest levels of DON and FB1 were detected in raw cereals and wholemeal flours. The highest levels of FB2 were detected in durum wheat pasta. A widespread DON contamination was found in baby and infant foods at levels varying from 0.007 to 0.166 μg g^-1.     

Lead levels in retail samples of Spanish infant formulae and their contribution to dietary intake of infants

Lead concentrations in 82 different types of infant formulae (cow's milk and soy based) marked in Spain were analysed by acid-microwave decomposition and anodic stripping voltammetry. Dietary lead intake from infant formula and tap water used for powder formula reconstitution were estimated in comparison with the provisional tolerable weekly intake (PTWI). Additionally, the influence of physical state (powder and ready-to-use formulae), the type of container used and the impact of the industrial process from different manufacturers on lead levels were evaluated. According to the results, lead exposure from drinking water was negligible with respect to formulae investigated; where soya formulae contributed the highest intake (58-73% PTWI), non-adapted starter and specialized formulae gave an moderate intake (31-42 and 26-37% PTWI, respectively), and, finally, pre-term, adapted starter and follow-up formulae provided the lowest lead intake (22-25, 22-26 and 16-22% PTWI, respectively). Based on the current state of knowledge about lead toxicity, manufacturers are called to make an additional effort in order to keep a maximum lead level at 20 µg l^-1 for all infant formulae, although it is recommendable that these formulations supply the upper limit (5 µg l^-1) of 'normal' human milk.     

Study of the mineral content of chocolate flavoured beverages

The concentrations of sodium, potassium, calcium, magnesium, sulfur and phosphorus (major elements) and iron, copper, zinc, selenium, molybdenum, chromium and manganese (minor elements) in chocolate-flavoured beverages from São Paulo State, Brazil, were determined. Different brands of this kind of beverage were analysed after digestion in a microwave oven and the quantification of the elements was done by inductively coupled argon plasma atomic emission spectrometry (ICP OES). Forty-four different samples from 13 different brands were studied, with significant differences in the contents of both major and minor elements formed between and within the different brands. However, it was possible to conclude that the samples analysed are good sources of K, Ca, Mg, S and P, and mainly those whose compositions included powdered milk. On the other hand, chocolate-flavoured beverages are shown to be poor sources of sodium. The data were examined by multivariate analysis with application of the method of pattern recognition (hierarchical grouping or cluster analysis).     

Patulin in domestic and imported apple-based drinks in Belgium: occurrence and exposure assessment

Apple-based beverages are regularly consumed by adults and children in Belgium. They are locally produced or imported from other countries. The apples used as starting material for these productions are frequently contaminated by mycotoxin-producing moulds and damaged during transport and handling. The current study was undertaken to investigate whether patulin (PAT) is present in the industrial or handicraft-made apple juices and ciders consumed by the Belgian population and to assess the population's exposure to this mycotoxin through apple-based drinks. Belgian (n = 29) and imported (14) apple juices as well as ciders (7) were assayed for PAT by high-performance liquid chromatography with ultraviolet light detection. PAT was detected respectively in 79, 86 and 43% of these tested samples. However, no contaminated sample exceeded the safe level of 50 µg PAT l^-1. Levels of PAT contamination were comparable in Belgian and imported juice samples. The overall mean PAT concentrations were 9.0 and 3.4 µg l^-1 for contaminated apple juices and ciders, respectively. This study also indicates that there was no statistically significant difference in the mean PAT contamination between clear (7.8 µg l^-1) and cloudy (10.7 µg l^-1) apple juices, as well as between handicraft-made apple juices (14.6 µg l^-1) and industrial ones (7.0 µg l^-1). On the basis of the mean results, a consumer exposure assessment indicates that a daily intake of 0.2 litres apple juice contributes to 45% of the provisional maximum tolerable daily intake for a child of 10 kg body weight.     

Determination of acrylamide in coffee and coffee products by GC-MS using an improved SPE clean-up

An improved gas chromatography-mass spectrometry (GC-MS) method to determine acrylamide (AA) in coffee and coffee products was developed. The method was based on two main purification steps: the first with ethanol and Carrez solutions in order to precipitate polysaccharides and proteins, respectively; and the second with a layered solid-phase extraction (SPE) column which proved to be efficient in the elimination of the main chromatographic interferences. The method is applicable to a wide range of coffee products. Twenty-six samples of different coffee products were analysed. The levels of AA were in the range 11.4-36.2 µg l^-1 for 'espresso coffee' and 200.8-229.4 µg l^-1 for coffee blends with cereals. The results indicate that the presence of cereals significantly increased the levels of AA.     

Determination of acrylamide in foods by pressurized fluid extraction and liquid chromatography-tandem mass spectrometry used for a survey of Spanish cereal-based foods

An automated and rapid method for the determination of acrylamide in different food products is presented. The method involves pressurized fluid extraction (PFE) of foods with acetonitrile and precipitation with Carrez reagents. The final extract is analysed by liquid chromatography coupled to electrospray ionization tandem mass spectrometry (ESI-MS-MS). The main parameters affecting the performance of ESI-MS-MS and PFE were optimized using a design of experiments approach. The limit of quantification of the method was 5 µg kg^-1, and recoveries from incurred samples ranged between 93 and 101%. The accuracy was evaluated using the reference test materials FAPAS T3002, T3005 and T3011. Using the optimized method, 62 food samples of potato chips, snacks, biscuits, breakfast cereals and crisp bread sampled from Valencia, Spain, supermarkets were surveyed for acrylamide levels. The levels were similar to those reported in the European Union and USA.     

Short communication: Discrimination between retail bovine milks with different fat contents using chemometrics and fatty acid profiling

We used a multivariate chemometric approach to differentiate or associate retail bovine milks with different fat contents and non-dairy beverages, using fatty acid profiles and statistical analysis. We collected samples of bovine milk (whole, semi-skim, and skim; n = 62) and non-dairy beverages (n = 27), and we analyzed them using gas-liquid chromatography. Principal component analysis of the fatty acid data yielded 3 significant principal components, which accounted for 72% of the total variance in the data set. Principal component 1 was related to saturated fatty acids (C4:0, C6:0, C8:0, C12:0, C14:0, C17:0, and C18:0) and monounsaturated fatty acids (C14:1 cis-9, C16:1 cis-9, C17:1 cis-9, and C18:1 trans-11); whole milk samples were clearly differentiated from the rest using this principal component. Principal component 2 differentiated semi-skim milk samples by n-3 fatty acid content (C20:3n-3, C20:5n-3, and C22:6n-3). Principal component 3 was related to C18:2 trans-9,trans-12 and C20:4n-6, and its lower scores were observed in skim milk and non-dairy beverages. A cluster analysis yielded 3 groups: group 1 consisted of only whole milk samples, group 2 was represented mainly by semi-skim milks, and group 3 included skim milk and non-dairy beverages. Overall, the present study showed that a multivariate chemometric approach is a useful tool for differentiating or associating retail bovine milks and non-dairy beverages using their fatty acid profile.     

Dietary intake of tin in Japan, and the effects on intake of canned food and beverage consumption

The study reported herein was initiated to examine dietary tin intake (Sn-D) in Japan to elucidate the possible effects of consumption of canned food (including beverages) on Sn-D, and to compare the intake among regions and between the two sexes in reference to the current provisional tolerable weekly intake and intake in other countries. Urinary tin levels (Sn-U) were also studied. Duplicate diet samples (24 h) together with records of food intake were collected in 1999-2004 from 111 adult residents in four areas of Japan. After exclusion of incomplete samples, 95 valid samples were subjected to determination of tin by inductively coupled plasma mass spectrometry (ICP-MS) after acid digestion. Among the 95 cases, 37 women additionally provided urine samples. Distribution of Sn-D was markedly skewed. Median Sn-D was 5.6 µg day^-1 for total subjects, which was about one-tenth of the values previously reported for the Japanese population; the difference was most probably attributable to the difference in the methods of determination. Consumption of canned foods led to a substantial increase in Sn-D. Thus, the median Sn-D for canned food consumers of 35.7 µg day^-1, was eight-fold higher than the median Sn-D for non-consumers of 4.5 µg day^-1. Sn-U (as corrected for creatinine concentration) distributed log-normally with a geometric mean of 2.0 µg (g cr)^-1. No effect of canned food consumption was evident on Sn-U. When compared internationally, Sn-D for the Japanese population was substantially lower than Sn-D for populations in other industrialized countries.     

Current incidence of N-nitrosodimethylamine in beers worldwide

A range of beers (n = 138) from 42 countries were analysed for the presence of N-nitrosodimethylamine (NDMA) using a chemiluminescence technique with a limit of quantification of 0.1 µg L^-1. The overwhelming majority of samples (79%) did not contain detectable NDMA, and only three samples exceeded 0.5 µg L^-1. No association was found between NDMA content and beer strength, type or geographical origin. It was noted that water can be a potential source of NDMA in beverages.     

Polycyclic aromatic hydrocarbons (PAHs) in meat products and estimated PAH intake by children and the general population in Estonia

The concentrations of benzo[a]pyrene and 11 other polycyclic aromatic hydrocarbons (PAHs) were analysed from 322 commercial, cured meat products and 14 home-grilled meat samples as part of the Estonian food safety monitoring programme during 2001-2005. The maximum acceptable concentration of 5 µg kg^-1 for benzo[a]pyrene was exceeded in 3.4% of samples. The highest PAH concentrations were detected in home-grilled pork samples. Using of disposable grilling unit resulted in 1.6 times higher PAH concentrations compared to the traditional wood-burning grill. The average intake of benzo[a]pyrene and sum of 12 PAHs from meat products was estimated for children (age 1-16 years) on the basis of an individual food consumption questionnaire and, for the general population, based on national food consumption data. The highest total PAH concentrations detected were 16 µg kg^-1 in smoked meat and ham, 19 µg kg^-1 in smoked sausage and 6.5 µg kg^-1 in smoked chicken samples. Since smoking and grilling are prevalent meat-cooking methods in Estonia, the impact of meat products is assessed to be significant in overall PAH intake.     

Wheat flour confectionery products as a source of inorganic nutrients: zinc and copper contents in hard biscuits

Cereal-based confectionery products being consumed through whole human life are considered mainly to be a source of carbohydrates, that is energy, although cereals are a rich source of minerals as well. To evaluate some hard biscuits produced in Croatia as a source of different trace elements in nutrition, in this study Zn and Cu contents were determined in classic wheat flour biscuits and in dietetic biscuits enriched with whole wheat grain flour or whole wheat grain grits, soya flour and skimmed milk. Zn was determined by flame atomic absorption spectrometry (AAS); Cu was determined by inductively coupled plasma atomic emission spectroscopy (ICP-AES). The results show that the Zn content in different kinds of biscuits ranges from 5.89 up to 17.64 mg/kg and the Cu content ranges from 1.15 up to 2.79 mg/kg depending on the type of wheat milling products and mineral content of other ingredients used. Enriched dietetic biscuits produced from wheat flour type 850 and whole wheat grain flour and/or soya flour and skimmed milk were almost 200% and 150% higher in Zn and Cu, respectively, in comparison to classic white wheat flour biscuits and can be considered as good sources of Zn and Cu in nutrition.     

Effects of Fusarium toxin-contaminated wheat and feed intake level on the biotransformation and carry-over of deoxynivalenol in dairy cows

An experiment was carried out to examine the effects of feeding Fusarium toxin-contaminated wheat (8.21 mg deoxynivalenol (DON) and 0.09 mg zearalenone (ZON) per kg dry matter) at different feed intake levels on the biotransformation and carry-over of DON in dairy cows. For this purpose, 14 ruminal and duodenal fistulated dairy cows were fed a diet containing 60% concentrate with a wheat portion of 55% (Fusarium toxin-contaminated wheat (mycotoxin period) or control wheat (control period)) and the ration was completed with maize- and grass silage (50 : 50) on a dry matter basis. Daily DON intakes ranged from 16.6 to 75.6 mg in the mycotoxin period at dry matter intakes of 5.6-20.5 kg. DON was almost completely biotransformed to de-epoxy DON (94-99%) independent of the DON/feed intake, and the flow of DON and de-epoxy DON at the duodenum related to DON intake ranged from 12 to 77% when the Fusarium toxin-contaminated wheat was fed. In the serum samples, de-epoxy DON was detected in the range of 4-28 ng ml^-1 in the mycotoxin period, while concentrations of DON were all below the detection limit. The daily excretion of DON and de-epoxy DON in the milk of cows fed the contaminated wheat varied between 1 and 10 µg and between 14 and 104 µg, respectively. The total carry-over rates as the ratio between the daily excretion of DON and de-epoxy DON into milk and DON intake were in the ranges of 0.0001-0.0002 and 0.0004-0.0024, respectively. Total carry-over rates of DON as DON and de-epoxy DON into the milk increased significantly with increasing milk yield. In the urine samples, de-epoxy DON was the predominant substance as compared with DON with a portion of the total DON plus de-epoxy DON concentration to 96% when the Fusarium toxin-contaminated wheat was fed, whereas the total residues of DON plus de-epoxy DON in faeces ranged between 2 and 18% of DON intake in the mycotoxin period. The degree of glucuronidation of de-epoxy DON was found to be approximately 100% in serum. From 33 to 80% of DON and from 73 to 92% of de-epoxy DON, and from 21 to 92% of DON and from 86 to 100% of de-epoxy DON were glucuronidated in the milk and urine, respectively. It is concluded that DON is very rapidly biotransformed to de-epoxy DON in the rumen and only negligible amounts of DON and de-epoxy DON were transmitted into the milk within the range of 5.6-20.5 kg day^-1 dry matter intake and milk yields (fat corrected milk) between 10 and 42 kg day^-1.     

Quantitative analysis of ochratoxin A in wine and beer using solid phase extraction and high performance liquid chromatography-fluorescence detection

A reliable and accurate method is described for the quantitative analysis of ochratoxin A (OTA) in wine and beer. The method involves the use of disposable non-polar polymeric and aminopropyl solid-phase extraction cartridges to isolate the mycotoxin from alcoholic beverages. Extracts were subsequently analysed using reverse-phase high-performance liquid chromatography-fluorescence detection with post column ammoniation to improve the limit of detection. The precision of the method determined at three levels in both wine and beer was less than 5% (RSD). Standard addition studies in both wine and beer showed that the recovery of OTA varied between 90 and 106% over a concentration range of 0.016-1.284 µg l^-1. The detection and quantification limits were shown to be better than 0.004 (S/N = 3) and 0.016 µg l^-1 (S/N = 10) for both beer and wine.     

Cadmium and mercury in cephalopod molluscs: Estimated weekly intake

Cadmium and mercury concentrations were measured in the flesh and hepatopancreas of different species of cephalopod molluscs (European squid, common octopus, curled octopus, horned octopus, pink cuttlefish, common cuttlefish) in order to establish whether the concentrations exceeded the maximum levels fixed by the European Commission. In hepatopancreas, the levels of cadmium were substantially higher than those in flesh (flesh 0.11-0.87 µg g^-1 wet weight, hepatopancreas 2.16-9.39 µg g^-1 wet weight), whilst the levels of mercury (flesh 0.13-0.55 µg g^-1 wet weight, hepatopancreas 0.23-0.79 µg g^-1) were approximately double those in flesh. Concentrations exceeding the maximum permitted limit of cadmium were found in 39.8 and 41.0% of common octopus and pink cuttlefish flesh, respectively. For mercury, concentrations above the limit were found only in octopuses, and precisely in 36.8, 50.0 and 20.0% of flesh samples of common, curled and horned octopus, respectively. In the hepatopancreas, concentrations of cadmium and mercury were above the proposed limits in all the samples examined. The estimated weekly intake of between 0.09 and 0.49 µg kg^-1 body weight for cadmium and between 0.05 and 0.24 µg kg^-1 body weight for mercury made only a small contribution to the provisional tolerable weekly intake (cadmium 1.3-7.0%, mercury 1.0-4.8%) set by the WHO.     

Assessment of hydroxymethylfurfural intake in the Spanish diet

Estimation of hydroxymethylfurfural (HMF) exposure of the Spanish population from heat-processed food was performed. HMF levels in Spanish foods were obtained from the data previously published in peer-reviewed papers; in addition, it was identified which food categories contributed significantly to HMF exposure. The potential HMF exposure was calculated for three different scenarios by using individual food intakes and the minimum (scenario 1), median (scenario 2) and maximum (scenario 3) values of analytical data on the HMF content in foods. A mean HMF intake of 10 mg day^?1 (corresponding to scenario 2) was obtained, which is only ten-fold lower than the tolerable daily intake. Coffee and bread are the most important food items that contribute nearly 85% to the total HMF daily exposure, although biscuits, breakfast cereals, beer, UHT milk and tomato products are also important for HMF exposure.     

Performance of new clean-up column for the determination of ochratoxin A in cereals and foodstuffs by HPLC-FLD

The performance of the newly developed Mycosep® 229 Ochra and Multisep® 229 Ochra clean-up columns for ochratoxin A (OTA) determination was evaluated. OTA was subsequently analysed using RP-HPLC with fluorescence detection. Recoveries for frequently contaminated commodities, like cereals, red wine, raisins and green coffee, were estimated. The recoveries obtained for the Mycosep 229 Ochra column were in the range from 87.9 ± 12.5% (n = 6) for wheat to 99.4 ± 2.7% (n = 24) for raisins. For Multisep 229 Ochra, recoveries from 76.5 ± 8.0% (n = 6) for barley to 86.4 ± 1.4% (n = 24) for raisins were achieved. Limits of detection for all matrices investigated (maize, wheat, rice, barley, raisins, green coffee beans, red wine) were in the range 0.4-2.4 μg kg^-1. The trueness of the method was tested using a certified reference material.     

A simple steady-state model for carry-over of aflatoxins from feed to cow's milk

A simple steady-state model is derived from two kinetic one-compartment models for the disposition of aflatoxin B₁ (AFB₁₎ and aflatoxin M₁ (AFM₁) in the lactating cow. The model relates daily intake of AFB₁ in feed of dairy cattle and the cow's lactation status to resulting concentrations of AFM₁ in milk. Moreover, assuming a linear relationship between the cow's lactation status and feed intake, the model relates daily milk production and AFB₁ concentration in total feed to AFM₁ levels in milk. The model explains similar experimental outcomes from different investigations into carry-over of aflatoxins from feed to milk. Although it is difficult to set a permanent limit for AFB₁ in feed, the European Union (EU) limit of 5 µg AFB₁ kg^-1 concentrate has proved, thus far, to be an appropriate level in preventing the EU limit of 0.05 µg AFM₁ kg^-1 milk being exceeded.     

Iodine content in commonly consumed food in Hong Kong and its changes due to cooking

Levels of iodine of foods found in Hong Kong were analysed in 271 samples from 11 groups, including (i) cereals and grain products, (ii) legumes and vegetables, (iii) meat and poultry, (iv) egg and egg products, (v) milk and milk products, (vi) fish, (vii) crustaceans and mollusks, (viii) non-alcoholic beverages, (ix) condiments and sauces, (x) sashimi and (xi) seaweeds. All food samples were analysed individually as purchased. The iodine in all samples ranged from undetectable to 2.9?g?kg^?1. Seaweeds, iodised salt, seafood, milk and milk products as well as egg and egg products were rich sources of iodine. To estimate the influence of cooking on iodine levels in foods, a total of 15 individual samples were analysed as raw and respective cooked food. The influence of cooking on the iodine level was minimal, except for boiling, as iodine dissolved into the soup.     

Annual variation of deoxynivalenol in Danish wheat flour 1998-2003 and estimated daily intake by the Danish population

The occurrence of deoxynivalenol (DON) in Danish wheat flour was studied during the period 1998-2003 by either capillary gas chromatography with electron capture detection and liquid chromatography coupled to an ion trap mass spectrophotometer. A total of 151 samples were collected from mills and the retail market in Denmark. Contamination levels varied considerably from year-to-year with the highest concentrations occurring in samples from the 2002 harvest with mean and median concentrations of 255 and 300 µg kg^-1, respectively. Compared to other harvest years, 2002 had the highest amount of precipitation around flowering time, i.e. from the end of June to the beginning of July covering weeks 25-27. The lowest average levels were found in samples from the 2001 harvest, where weeks 25-27 were dry compared with other harvest years. The highest value (705 µg kg^-1) was obtained in a flour sample from the 2002 harvest, but none of the tested samples exceeded the maximum limit of 750 µg kg^-1, which has been recently introduced by the European Commission for DON in flour used as raw materials in food products. Calculation of chronic or usual intake by a deterministic approach showed that intake did not exceed the TDI of 1 µg kg^-1 bw day^-1 either for the whole population or for children. A probabilistic approach also showed that intake in general was below the TDI, but intake for children in the 99% percentile amounted to more than 75% of the TDI. The highest intake is calculated to be 2.5 µg kg^-1 bw day^-1.