Objectives: For deep carious lesions, selective carious tissue removal (leaving soft dentin close to the pulp) is suggested. Afterward, different restoration materials, such as resin composites or glass hybrids (GHs), can be placed. Many dentists also apply setting or non-setting calcium hydroxide liners before restoration. We compared margin integrity and susceptibility for secondary caries in differently restored premolars in vitro. Methods: In 48 extracted human premolars, artificial residual lesions were induced on pulpo-axial walls of standardized cavities. Teeth were restored using a GH (Equia Forte) or adhesively placed resin composite restoration (OptiBond FL and Tetric EvoCeram) without any liner (RC), resin composite restoration with a non-setting calcium hydroxide liner (RC_NCH), or resin composite restoration with a setting calcium hydroxide liner (RC_SCH). After thermomechanical cycling, groups (n=12) were compared regarding their gingivocervical margin integrity (proportion of irregularities, microgaps, gaps > 5 mu m, overhangs). Teeth were then submitted to a continuous culture Lactobacillus rhamnosus biofilm model. After 14 days, bacterial numbers in biofilms, along tooth-restoration margins and mineral loss (DZ) of secondary lesions, were determined. Results: GH and RC_NCH showed significantly higher proportions of irregularities than RC RC_SCH (p < 0.05/Mann-Whitney). GH also showed significantly more gaps than alternative restorations (p < 0.05). Bacterial numbers and Delta Z did not differ significantly between groups (p > 0.05). Conclusions: GH and composites lined with non-setting calcium hydroxide showed reduced margin integrity compared with non-lined composites or composites lined with setting calcium hydroxide. This did not increase susceptibility for secondary caries.