BackgroundThe change in mechanical properties of femoral cortical bone tissue surrounding the stem of the hip endoprosthesis is one of the causes of implant instability. We present an analysis used to determine the best conditions for long-term functioning of the bone–implant system, which will lead to improvement of treatment results. MethodsIn the present paper, a finite element method coupled with a bone remodelling model is used to evaluate how different three-dimensional prosthesis models influence distribution of the density of bone tissue. The remodelling process begins after the density field is obtained from a computed tomography scan. Then, an isotropic Stanford model is employed to solve the bone remodelling process and verify bone tissue adaptation in relation to different prosthesis models. FindingsThe study results show that the long-stem models tend not to transmit loads to proximal regions of bone, which causes the stress-shielding effect. Short stems or application in the calcar region provide a favourable environment for transfer of loads to the proximal region, which allows for maintenance of bone density and, in some cases, for a positive variation, which causes absence of the aseptic loosening of an implant. In the case of hip resurfacing, bone mineral density changes slightly and is closest to an intact femur. InterpretationInstallation of an implant modifies density distribution and stress field in the bone. Thus, bone tissue is stimulated in a different way than before total hip replacement, which evidences Wolff's law, according to which bone tissue adapts itself to the loads imposed on it. The results suggest that potential stress shielding in the proximal femur and cortical hypertrophy in the distal femur may, in part, be reduced through the use of shorter stems, instead of long ones, provided stem fixation is adequate.