Certain mycotoxins may be present in plant materials as their glucosides. The question is whether these glucosides may be hydrolysed into their parent compounds in the gastro-intestinal tract (GI-tract), thus increasing the exposure. Therefore, the potential hydrolysis of deoxynivalenol-3-ß-glucoside (DON-3G) to deoxynivalenol (DON) was assessed in two in vitro models representing the human upper GI-tract (mouth, stomach and small intestine). In a fed digestion model, there was no evidence of release of DON from DON-3G, spiked at a level of 2,778 µg DON- 3G/kg food. This shows that the conditions in the GI-tract do not result in hydrolysis of this glucoside into the original mycotoxin. The absorption and transformation of DON-3G in the small intestine was assessed in an in vitro model with human Caco-2 cells in a Transwell system. No evidence was found for the transformation of DON-3G to DON by the Caco-2 cells in both the apical or basolateral side in 24 hours (cells were exposed to 2.4 nmol DON- 3G/ml medium). However, when DON itself was added to the apical side an amount of 23% of the spiked DON was detected in the basolateral side after 24 hours (cells were exposed to 2.3 nmol/ml medium). In conclusion, no evidence was found in the in vitro experiments for significant elevated exposure of humans to DON, since DON- 3G was not hydrolysed to DON in the digestion model representing the upper part of the GI-tract and DON-3G was not hydrolysed to DON by the intestinal epithelial Caco-2 cells. It was shown that bioavailability of DON-3G in humans may be low as compared to DON since Caco-2 cells did not absorb DON-3G, in contrast to DON.