The aim of the study was to determine the retentive strength of monolithic all-ceramic crowns cemented on titanium implant abutments. 225 crowns (75 crowns each of Mark II, Empress CAD, and e.max CAD) were milled using a CAD/ CAM system. The crowns were cemented onto sandblasted titanium implant abutments using five luting agents (Multilink Implant, Variolink II, RelyX Unicem, Fujicem, and Panavia 2.0). After thermocycling, the crowns were removed using a universal testing machine. The location of luting agent residue on the abutment and inner crown surfaces was evaluated. Statistical analysis was performed uisng ANOVA with the Bonferroni correction. In comparing the luting agents, no significant difference was found between Vita Mark II and Empress CAD. For e.max CAD, the luting agent RelyX Unicem had a significantly higher mean retentive strength than did Multilink Implant (p = 0.003) or Panavia 2.0 (p = 0.001). In comparing the ceramic materials, e.max CAD showed significantly higher pull-off strengths than the other two ceramic materials when the luting agents RelyX Unicem and Fujicem were used (all p 0.001). The residues of nearly all luting agents were located entirely or almost entirely (75%-100%) on the inner crown surfaces of all ceramic materials, except for the luting agent GC Fujicem, which left more luting agent residue (0%-75%) on the abutment surfaces of all ceramic materials. In comparing the five luting agents, significant differences in the resulting retentive strength were only found for the ceramic material e.max CAD. The other ceramic materials did not show significant differences in retentive strength, independent of the luting agent. Keywords: implant, ceramic, retentive strength, CAD/CAM, luting agent, cement