Finite element analysis (FEA) is nowadays a pivotal tool in orthopaedic research for personalized virtual surgery planning. Despite its widespread use, a comprehensive evaluation of the effect of boundary conditions on the simulation of physiological mechanics in implanted bone is currently lacking. This study assesses the impact of boundary conditions and femur geometry on predicted femur mechanics. It focuses on an isolated implanted femur, partially imaged, from a paediatric patient with femoral varus who underwent a proximal femoral osteotomy. By employing FEA of the femur under motion with loading scenarios informed by personalized neuromusculoskeletal modelling, this study evaluated implant and bone mechanics across three femur model configurations (full-femur, proximal half-femur, and distally synthesized full-femur) with two boundary condition approaches (biomechanical and fixed distal). The biomechanical boundary condition was validated against the gold standard inertia relief method for the intact femur and thereafter exploited as the benchmark against the other implanted femur model configurations. The distally synthesized full-femur with biomechanical boundary conditions performed best and closely predicted bone-implant micromotion (R2=0.99, nRMSE=0.3%), risk of implant yield (<1% variance from the benchmark model), and interfragmentary movement (R2=1, nRMSE=6%). The half-femur model with biomechanical boundary conditions overpredicted the risk of yield and interfragmentary movements by 17% and 15.8%, respectively. The fixed distal constraint method significantly overestimated the risk of implant yield in both half and synthesized full-femur models by 157% and 170%, respectively. These findings underscore the critical importance of selecting appropriate boundary conditions in the FEA of implanted femur models and advocate for the synthesis of the missing portion of the femur coupled with the biomechanical boundary conditions for more accurate predictions of bone and implant mechanics. Such insights are expected to enhance the physiological plausibility and reliability of orthopaedic research and clinical practices, especially when managing proximal femoral osteotomies.