The presence, dissolution, agglomeration state, and release of materials in the nano-size range from food containing engineered nanoparticles during human digestion is a key question for the safety assessment of these materials. We used an in vitro model to mimic the human digestion. Food products subjected to in vitro digestion included (i) hot water, (ii) coffee with powdered creamer, (iii) instant soup, and (iv) pancake which either contained silica as the food additive E551, or to which a form of synthetic amorphous silica or 32 nm SiO(2) particles were added. The results showed that, in the mouth stage of the digestion, nano-sized silica particles with a size range of 5-50 and 50-500 nm were present in food products containing E551 or added synthetic amorphous silica. However, during the successive gastric digestion stage, this nano-sized silica was no longer present for the food matrices coffee and instant soup, while low amounts were found for pancakes. Additional experiments showed that the absence of nano-sized silica in the gastric stage can be contributed to an effect of low pH combined with high electrolyte concentrations in the gastric digestion stage. Large silica agglomerates are formed under these conditions as determined by DLS and SEM experiments and explained theoretically by the extended DLVO theory. Importantly, in the subsequent intestinal digestion stage, the nano-sized silica particles reappeared again, even in amounts higher than in the saliva (mouth) digestion stage. These findings suggest that, upon consumption of foods containing E551, the gut epithelium is most likely exposed to nano-sized silica.