Background: Few studies have focused on treatment following failed total ankle replacement. The purpose of this study was to report the outcomes of patients undergoing revision total ankle replacement and to propose a talar component subsidence grading system that may be helpful in making decisions regarding how to revise failed total ankle replacements in the future. Methods: A retrospective review was performed of fifty-three patients who underwent revision total ankle replacement and had been followed for a minimum of two years. Patients were assessed radiographically and with outcome scores. The rates of conversion to amputation or fusion were also assessed. Results: The mean follow-up period was 49.1 months after the revision arthroplasty. The average time from primary total ankle replacement to revision was fifty-one months. Forty-one of the fifty-three patients (77%) were available for follow-up. The revision arthroplasty had been converted to an arthrodesis in five of the forty-one patients, and two additional patients had undergone amputation. The most common indication for revision total ankle replacement was talar subsidence (63%; twenty-six of forty-one). Twenty-two patients (54%) had a subtalar arthrodesis performed at the time of the revision arthroplasty, with nineteen of those having a custom-designed long-stem talar component placed simultaneously. The mean radiographic measurements of component position did not change significantly postoperatively. The mean postoperative scores for the thirty-four patients with a retained total ankle replacement were: 4.4 of 10 possible points on a visual analog pain scale (VAS), 65 of 100 possible points on the American Orthopaedic Foot & Ankle Society (AOFAS) hindfoot scale, 93.5 of 100 possible points on the Short-Form 12 (SF-12), 137.9 of 204 possible points on the Revised Foot Function Index (FFI-R), and 64 of 180 possible points on the Ankle Osteoarthritis Scale (AOS). The mean arc of motion radiographically was 18 preoperatively and 23 postoperatively, with all improvement occurring in plantar flexion. A lesser amount of preoperative talar subsidence was a significant predictor of a good outcome based on the AOFAS hindfoot score (p < 0.03) and the AOS (p < 0.01) score. Conclusions: Revision arthroplasty may be considered as an alternative to arthrodesis when treating patients with a failed Agility total ankle implant.
