The study aimed to define the in vitro secondary caries inhibiting potential of restorative materials currently used in dental practice. Class V restorations were prepared in extracted human third molars and immersed in a demineralizing solution (lactic acid, pH 4.5) at 37 degrees C for 2 days to simulate secondary caries formation. The bonding and the restorative systems tested in the study were: Scotchbond 1+Z 250 (Group A), Scotchbond 1+F 2000 (Group B), ABF+APX (Group C), ABF+F2000 (Group D). Perimarginal dentine, immediately close to the margin of the restoration, and exposed dentine, at approximately 0.5 mm from the margins of the restoration, after exposure to the acid solution, were investigated; protected dentine, at approximately 4 mm from the margin in a varnish-covered area, was analysed as control. Polarized light microscopy and contact transverse microradiography (TMR) were employed. The output parameters were lesion shape and size (depth in microm) of the exposed dentine, dentine mineral volume%, and integrated mineral loss (Delta Z, in %volmicrom) of the lesions. Compomers (Groups B and D) showed a thinner demineralization of the outer lesions, a less demineralization along the perimarginal dentine (inner lesion) and more caries inhibition zones or CIZs (Delta Z positive values) compared to composites (Groups A and C). In conclusion, Groups B and D materials seemed to partially counteract the marginal demineralization induced by an acid solution and favourably influence the formation of CIZs along the restorations. On the contrary, composites did not show a protective effect, probably due to an insufficient marginal seal and the lack of fluoride release.