Colloidal silica is widely used in topical products for its effect on the rheology of the formulations. It is considered a non-toxic and non-irritant excipient and it is included in the European and US Pharmacopoeias. However, in vitro investigations have shown a pro-inflammatory stimulation of endothelial cells induced by exposure to silica particles of nano-size. In the light of the recent concern about the potential toxicological risk associated with topically applied nanoparticles, aim of the present study was to investigate the in vivo skin penetration profile of commercial colloidal silica from a model cream (oil-in-water emulsion) preparation. Colloidal silica (Aerosil 200; BET specific surface area, 200 m2/g) was incorporated in the water phase of the test emulsion, after dispersion by sonication (15 min). The emulsion was applied at a level of 4 mg/cm2 to the forearm of the volunteers, in accordance with the local Ethics Committee requirements and the percutaneous permeation of the silica nanoparticles was assessed by the tape-stripping technique. The tapes were collected separately and the silica distribution in each group of tapes was determined by inductively coupled plasma quadrupole mass spectrometry assay of the silicon content of the strips. Total recovery of colloidal silica (sum of silicon unabsorbed and transferred to the strips) was 97.8%. Negligible amounts (1.5 μg/cm2 ± 0.2) of silicon were detected in the tapes from the stratum corneum exposed to the control formulation without silica. On the other hand, the majority (58.9%) of the applied silica nanoparticles was found to diffuse up to the intermediate region of the human horny layer, with only 2.1% detected in the deepest strip. Since the silica particles used in the present investigation were mainly confined to the upper stratum corneum layers, potential toxicological risk should be limited.