Purpose: This study compared the removal torque values (RTVs) of different abutments (straight, angled, and gold premachined UCLA-type) in external- and internal-hex implants after dynamic cyclic loading with the clinical situation of the anterior maxilla simulated. Materials and Methods: An ideal cast of a maxilla with a missing right central incisor was fabricated in dental stone, and an implant analog was embedded in this model at a 15-degree angle labial to the long axis of the left central incisor. Thirty external-hex and 30 internal-hex implants were used. A total of 10 straight abutments, 10 angled abutments, and 10 gold premachined UCLA-type abutments of each system and 60 abutment screws were tested. Initial RTVs were measured after each assembly was tightened to 30 Ncm. Straight abutments and angled abutments were prepared and gold-premachined UCLA-type abutments were waxed and cast with low-fusing gold alloy for the central incisor. RTVs were then measured again. After each assembly was tightened, a metal crown was temporarily cemented. After cyclic loading of 20 to 200 N was applied 1 million times, RTVs were measured for a third time. Statistical analysis (α = .05) was performed to evaluate the results. Results: The angled abutment group showed significantly higher RTVs than the straight abutment and gold premachined UCLA-type abutment groups in external-hex implants. However, no significant difference in RTVs was found among abutments in internal-hex implants. The time of analysis of RTV was found to significantly influence mean RTVs. Mean RTVs of external- and internal-hex implants showed significant differences. Conclusions: Within the limitations of this study, there were significant differences in RTVs among different abutment groups in external-hex implants. There were no significant differences in RTVs among different abutment groups in internal-hex implants. Keywords: abutment angulation, anterior maxilla, dental implant, dynamic cyclic loading, prosthetic screw, screw loosening