PURPOSE. Limited data is available regarding the differences for possible microleakage problems and fitting accuracy of zirconia versus titanium abutments with various connection designs. The purpose of this in vitro study was to investigate the effect of connection design and abutment material on the sealing capability and fitting accuracy of abutments. MATERIALS AND METHODS. A total of 42 abutments with different connection designs [internal conical (IC), internal tri-channel (IT), and external hexagonal (EH)] and abutment materials [titanium (Ti) and zirconia (Zr)] were evaluated. The inner parts of implants were inoculated with 0.7 mu L of polymicrobial culture (P gingiyalis, T forsythia, T denticola and F. nucleatum) and connected with their respective abutments under sterile conditions. The penetration of bacteria into the surrounding media was assessed by the visual evaluation of turbidity at each time point and the number of colony forming units (CFUs) was counted. The marginal gap at the implant- abutment interface (IAI) was measured by scanning electron microscope. The data sets were statistically analyzed using Kruskal-Wallis followed by Mann-Whitney U tests with the Bonferroni-Holm correction (alpha=.05). RESULTS. Statistically significant difference was found among the groups based on the results of leaked colonies (P<.05). The EH-Ti group characterized by an external hexagonal connection were less resistant to bacterial leakage than the groups EH-Zr, IT-Zr, IT-Ti, IC-Zr, and IC-Ti (P<.05). The marginal misfit (in pm) of the groups were in the range of 2.7-4.0 (IC-Zr), 1.8-5.3 (IC-Ti), 6.5-17.1 (IT-Zr), 5.4-12.0 (IT-Ti), 16.8-22.7 (EH-Zr), and 10.3-15.4 (EH-Ti). CONCLUSION. The sealing capability and marginal fit of abutments were affected by the type of abutment material and connection design.