This study evaluated the effect of adhesives on the inhibition of secondary caries around compomer restorations in vitro. Two adhesive systems with a Bis-GMA resin, Scotch bond Multi-purpose (MP) and Single Bond (SB), and one adhesive system with no Bis-GMA resin, F2000 compomer primer/adhesive (PA), were used prior to placement of the compomer (F2000), and non-fluoride releasing resin composite (Z100) was used as a control. Class V cavities prepared on extracted human premolars were restored with various combinations of materials: F2000/MP, F2000/SB, F2000/PA, Z100/MP, Z100/SB and Z100/PA. The restored teeth were incubated in bacterial medium containing sucrose with Streptococcus mutans for two weeks after storage for 14 days. On microradiographs, the radio-opaque layers adjacent to the F2000 restorations were thick and clear, while the layers in the Z100 restorations were unclear. In the F2000 restorations, the mean thickness of the radio-opaque layers in the PA group was significantly greater than that of the MP and SB groups. In fluoride-releasing measurement, F2000 coated with PA showed a significantly higher amount of fluoride release than MP and SB, and no significant difference in the amount of fluoride release from uncoated F2000. These results indicated that applying an adhesive without Bis-GMA resin to compomer restoration has no suppressive effect on the fluoride release from compomer and might be beneficial for inhibiting secondary caries in vitro.