Effect of Food Processing on Mycotoxin Levels
Annalisa De Girolamo and Michele Solfrizzo
Thermal processing such as baking, extrusion cooking and roasting are technologies used by the food industry to produce processed foods. These processes together with milling are treatments that may affect mycotoxins stability causing a destruction, change, removal or redistribution of mycotoxins in the processed food.
Fumonisins
Although fumonisins are fairly heat stable, several studies showed that levels of fumonisin B1 (FB1) recovered in the final product decreased as the temperature or heating time increased. Screw speed and barrel temperature significantly affected the extent of FB1 reduction in the extruded grits; in particular, the reduction of FB1 in corn grits spiked with 5 µg/g FB1 increased from 34 to 95% as barrel temperature increased from 140 to 200 Deg C and screw speed decreased from 160 to 40 rpm. Acceptable properties of the extruded product were obtained at temperatures of 160-200 Deg C and speeds of 120-160 rpm with a reduction of about 46-76% of FB1 level (1).
The addition of glucose further reduced the fumonisin level during extrusion; in particular, losses of FB1 up to 92% were observed when corn grits artificially contaminated with 5 µg/g FB1 were extruded (160 Deg C and screw speeds of 40 rpm) in presence of 10% of glucose (2). Losses of fumonisins seem to be also related to the formation of reaction products especially in presence of reduced sugars, therefore the toxicity of these compounds need to be assessed.
Accumulation of Fumonisins in germ and bran during milling of contaminated corn has been recently confirmed. Analysis of corn meal, corn grits, corn flour, germ and bran, originating from dry-milled corn naturally contaminated with 9.7 µg/g of fumonisins showed that levels of fumonisins in germ and bran were 3-fold higher than the whole corn, 13-fold higher than the corn flour and 29-fold higher than the corn meal and corn grits (3). These results suggest that germ and bran fractions are at high risk of fumonisin contamination and should be constantly monitored.
Aflatoxins
Aflatoxins are relatively heat-stable and are not completely destroyed when thermally treated to produce food or feed. Traditional roasting of green coffee (naturally contaminated with 6.00 µg/kg AFB1) at 180 Deg C for 10 min reduced aflatoxin B1 (AFB1) level by 51%. No reduction of AFB1 was observed during the decaffeination process. (4).
Extrusion temperature influenced AFB1 content in cotton seed naturally contaminated with 339 µg/kg AFB1; in particular at the extrusion temperature of 104 and 160 Deg C, AFB1 reductions of 46 and 65%, respectively were observed. Aflatoxin levels were reduced by an additional 55% when the cottonseed was extruded 4 times compared to only once (5).
Studies performed on aflatoxin M1 (AFM1) reduction during processing, ripening and storage of cheese agree that the toxin is not degraded during conversion of AFM1-contaminated milk into pressed curd and that removal of whey during cheese-making eliminates appreciable amount of the toxin. The variable information on AFM1 partition between curd and whey is probably due to the type of cheese analysed, the type of cheese-making process applied, the degree of milk contamination and the analytical method employed. The distribution of AFM1 between curd and whey was studied during the preparation of Telemes cheese starting from milk contaminated with 0.045 – 0.088 µg/L AFM1. The percentages of AFM1 in curd and whey with respect to the initial AFM1 concentration in milk were 54–60% and 44-40%, respectively (6). Distribution and stability of AFM1 during production and storage of yoghurt has been recently studied. The percentage losses of the initial amount of AFM1 in milk were estimated at about 13 and 22% by the end of the fermentation, and 16 and 34% by the end of storage for yoghurts with pHs 4.6 and 4.0, respectively (7).
Deoxynivalenol
Wet and dry milling is a common method of processing that has been shown to reduce the level of deoxynivalenol (DON) in corn and wheat. Other treatments such as cleaning, baking, extrusion and cooking, combined with milling further reduces DON levels. Reduction of DON during durum wheat processing and spaghetti cooking has been recently studied (8). With respect to the uncleaned wheat, containing 0.3 – 13.2 µg/g of toxin, a DON reduction of 23% in cleaned wheat, 36% in fine middlings, 63% in semolina and 67% in spaghetti was observed. Average DON levels in the screenings and bran fractions were 4.1 and 1.6 fold higher than the uncleaned wheat, respectively. About 55% of DON present in uncooked spaghetti was released into cooking water during cooking of spaghetti. The retention level of DON from grains on the market to cooked pasta on the plate has been assessed at 25% or less.
Ochratoxin A
Results of an EU funded project on the fate of ochratoxin A (OTA) during malting showed that an overall reduction of 79% was achieved during beer production from malt contaminated with 470 µg/g of OTA; in particular mashing accounted for 46%, boiling, wort clarification and cooling for 3%, fermentation for 15% and stabilization and filtration for 15%.
Several investigations on the effect of the roasting process on the OTA levels in coffee beans reported an OTA reduction ranging from 0 to 100%. A possible explanation of these inconsistent and contradictory results could be the heterogeneity of natural coffee bean contamination, initial OTA levels, differences in the method of introducing the toxin into the ground beans, analytical method performances and roasting conditions. The influence of different roasting conditions (450 Deg C for 6, 7 and 9 min) on OTA content of 4 naturally contaminated (1.99-29.30 µg/kg) green coffee samples have been investigated (9). The reduction of OTA ranged from 0-75% during light roasting, from 19-91% during medium roasting and from 91-97% during dark roasting, depending on initial OTA content. Since chaff removal contributes only for 2%, thermal destruction and initial OTA content were the principal factors responsible for OTA reduction during roasting of green coffee.
The fate of OTA in the processing of whole wheat grains contaminated with 5-40 µg/kg OTA during milling and bread production has been reviewed (10). Levels of OTA in the cleanings, scourings, bran and offal fractions were much higher than in the whole wheat, but lower in the white flour. An overall reduction of about 75% was obtained in white bread using a combination of cleaning, scouring and a removal of the bran and offal fractions, whereas a maximum overall reduction of 40% was observed during production of wholemeal bread. This indicates that the baking process itself is relatively ineffective at destroying significant amount of OTA.
Further information on this subject can be found in the first fact sheet and in the training course.
References
1. Katta SK, Jackson LS, Sumner SS, Hanna MA and Bullerman LB, 1999. Cereal Chem. 76, 16-20.
2. Castelo MM, Jackson LS, Hanna MA, Reynolds BH and Bullerman LB, 2001. Food Chem. Toxicol., 66, 416-421.
3. Broggi LE, Resnik SL, Pacin AM, Gonzalez HHL, Cano G and Taglieri D, 2002. Food Addit. Contam., 19, 465-469.
4. Soliman KM, 2002. J. Agric. Food Chem., 50, 7477-7481.
5. Buser MD and Abbas HK, 2002. J. Agric. Food Chem., 50, 2556-2559.
6. Govaris A, Roussi V, Koidis PA and Botsoglou NA, 2001. Food Addit. Contam., 18, 437-443.
7. Govaris A, Roussi V, Koidis PA and Botsoglou NA, 2002. Food Addit. Contam., 19, 1043-1050.
8. Visconti A, Haidukowski EM, Pacsale M and Silvestri M, 2003 (Toxicology Letters, submitted).
9. Romani S, Pinnavaia GG and Dalla Rosa M, 2003. J. Agric. Food Chem., 51, 5168-5171.
10. Scudamore KA, Banks J and MacDonald SJ, 2003. Food Addit. Contam., 20, 1153-1163.