Pre-clinical testing of hip endoprostheses is a mandatory requirement before clinical release. Inadequate loading conditions may lead to lower elastic and plastic interface movements than those occurring post-operatively in vivo. This study investigated the influence of patient activity on the primary stability of cementless prostheses with a special emphasis on active simulation of muscle forces. A loading setup, based on validated musculo-skeletal analyses, was used to generate the hip contact force during walking and stair climbing by transmitting muscle forces through the femur. In addition, a loading configuration which only generated the hip contact force occurring during stair climbing at the prosthesis head was simulated. CLS prostheses were implanted in 18 composite femora and subjected to cyclical loading. The relative micro-movements at the bone-prosthesis interface were determined and appeared to be extremely sensitive to the specific patient activity. Compared to walking, stair climbing generated higher micro-movements, with pronounced axial and rotational components. Stair climbing with the femur loaded by the resultant hip contact force only exhibited a characteristic valgus tilt of the stem with significantly lower interface micro-movements than under active simulation of muscle forces. The analyses suggest that stair climbing induced the highest mechanical instability at the bone-prosthesis interface, a level which may compromise the necessary osseointegration process. Active simulation of muscle forces considerably affects the primary stability of cementless hip endoprostheses. Pre-clinical in vitro tests should therefore simulate stair climbing and include muscle activity in the assessment of initial implant stability, otherwise micro-movements may be underestimated and the primary stability overestimated.