Purpose The aim of this work was to develop a computational biomechanical study to compare the performance of tibial trays with different offsets for a total knee arthroplasty. The goal was to investigate whether the offset tibial tray shifts the bone stress distribution, influencing the clinical outcome. Methods Three geometric models were developed for the intact tibia bone: one considering a standard tibia case and the other two reproducing tibias with a medial or a lateral offset of the metaphysis. Appropriate prosthetic components were assembled in the bone for the aforementioned cases. The finite element method was used to obtain the mechanical stress distribution for the models, and the stress shielding effect due to the prosthesis was analysed. Results The obtained results revealed that the offset cases are subjected to higher stresses than the standard case. These values can be two times superior to the ones verified in a standard case. The stress shielding effect was confirmed along all the analysed paths, except near the stem's end in some areas. Conclusion The higher stresses registered can originate lower clinical outcomes in the offset cases. These findings can be an important beginning to understand whether better bone stress distribution could be achieved in deformity correction with associated osteotomies instead of offsetting.