A prospective, comparative study of two Total Knee Replacement systems was undertaken to compare the traditional end point of revision of the prosthesis with three other end points based on the British Orthopaedic Association Knee Assessment Protocol. These were an unacceptable pain score plus revision, Knee Score of 60% or less plus revision, and a fall to below the preoperative Knee Score plus revision. 182 Knees entered the trial (74 St Leger and 108 Miller-Galante 11 prostheses), and all were assessed for Knee Score pre-operatively. Patients were subsequently assessed for Knee Score at 6 months, 1 year and annually thereafter. 63 implants had reached the five year stage at the time of this study.

The results indicate that different patients are designated as ‘failures’ depending on the end point selected, and that different survivorship figures can therefore be calculated with up to 6% variation depending on the end point used. In addition, the most successful prosthesis varies with the end point selected. Although the 95% confidence intervals in this study did not permit us to draw any firm conclusions, we suggest that further investigation into the definition of failure in survivorship analysis is necessary in order to reflect the performance as well as the longevity of Total Knee Replacements.