End-stage osteoarthritis is characterised by pain and reduced physical function, for which total knee arthroplasty (TKA) is recognised to be a highly effective procedure. Post-operative outcome and resultant function however is variable. Many factors are thought to influence outcome; in particular quadriceps muscle strength is one of the strongest predictors of the patient’s ability to perform functional tasks. Muscle atrophy has been shown to account for only a third of the variance in muscle power, the remainder is currently unexplained. In this thesis it is hypothesised that physiological and mechanical factors will affect muscle power post TKA. A new design of prosthesis with an axis of rotation of the knee based on new kinematical observations has been suggested to confer a mechanical advantage to the knee extensor mechanism by lengthening its moment arm, and thus reducing the muscular effort required to extend the knee, however this has not as yet been clinically demonstrated. A strong extensor mechanism is recognised as being paramount to the patients return to functional activity following TKA, but there has been no consideration as to the mechanisms how and to what extent the muscle tissue actually recovers. It is known that muscle satellite cells are essential for the regeneration of skeletal muscle and that these cells are activated following damage, but these have not been considered in relation to recovery from orthopaedic procedures. It is hypothesised that the number of satellite cells in the extensor mechanism will vary in the patient population and will influence muscle recovery. A double blind randomised controlled trial of 212 TKA patients was conducted to compare the new implant design with a traditional model. Patient outcome was assessed at four points over a one year period. The new implant was superior in measures of knee flexion, lower limb power output and by patient report questionnaire (Oxford Knee Score) Two-way ANOVA, p = <0.001 in all cases. Extensor mechanism power was significantly increased between all four assessment points in the new implant group, the control group demonstrating change between the second and third assessment only (p= <0.001). Analysis of the outcome assessments used demonstrated a changing relationship between function and patient report of that function. Regression models demonstrated that patient report of function became more consistent with direct functional assessment as the influence of pain diminished post-operatively. A hierarchical model is presented that highlights the limitation of patient report data in isolation. Muscle satellite cells were isolated from biopsies of the quadriceps muscle of 18 patients at the time of surgery and counted by an immunofluorescent staining technique. The number of satellite cells detected accounted for a third of the postoperative variance in power output (R2 = 36.6%). This was confirmed in another cohort of 11 patients with a more sensitive qPCR technique. It was further found that the activated satellite cells accounted for around two thirds of the change in postoperative power output (R2 = 66.7%). In conclusion, both mechanical and physiological factors have a significant effect on muscle power post total knee arthroplasty.