Objectives: Evaluate the fatigue resistance and failure mode of novel-design implant restorations made of histoanatomic computer-assisted design/computer-assisted manufacturing (CAD/CAM) bilaminar bonded assemblies. Material and methods: 60 screw-retained implant restorations were fabricated. Monolithic restorations were used as a control group (15 lithium disilicate: group CE; and 15 nanofilled composite resin: group CL), and compared with bilaminar restorations (15 dentin-shaped lithium disilicate mesostructures with a bonded nanofilled composite resin veneer: group CEL; and 15 dentin-shaped nanofilled composite resin mesostructures with a bonded lithium disilicate veneer: group CLE). All monolithic and bilaminar restorations were assembled and bonded to a customized metal implant abutment. Cyclic isometric chewing (5 Hz) at a 30-degree angle was simulated, starting with 5,000 cycles at a load of 150 N, followed by 20,000 cycles with increments of 50 N. Samples were loaded until fracture or to a maximum of 160,000 cycles. The groups were compared using the Life Table survival analysis (logrank test at P = 0.05; post hoc tests at P = 0.008). Results: In the CL group, restorations failed at an average load of 347.39 N (98.361 cycles), and in the CLE group at an average load of 313.20 N (83.105 cycles), and none of the specimens withstood all 160,000 load cycles. In the CE group, restorations failed at an average load of 381.47 N (119.115 cycles), and in the CEL group at an average load of 415.20 N (132.873 cycles), with survival rates of 26% and 33%, respectively. Post hoc tests from the load step data showed higher fatigue resistance of CEL compared with CLE (P = 0.003). Conclusion: Lithium disilicate monolithic restorations and mesostructures with nanofilled composite resin veneer presented higher survival rates when compared with nanofilled composite resin restorations.