Revision patellofemoral arthroplasty: three- to seven-year follow-up.

In this series, a failed first-generation patellofemoral arthroplasty was revised to a second-generation prosthesis. Fourteen knees were prospectively followed up for a mean of 60 months. Primary procedure failure was due to component malposition, subluxation, polyethylene wear, or overstuffing. Mean Bristol knee scores improved from 58 (range, 36-86) to 79 (range, 38-100) (P < .001). Mild femorotibial arthritis (Ahlbach stage I) was present in 5 knees and predicted a poorer outcome. At most recent follow-up, there was no evidence of wear, loosening, or subluxation. Significant improvement can be obtained when revising the failed patellofemoral arthroplasty, provided there is no femorotibial arthritis.