Outcomes of Arthrodesis for the Revision of Failed Total Ankle Arthroplasty

Research Type: Level 3 - Retrospective cohort study, Case-control study, Meta-analysis of Level 3 studies Introduction/Purpose: As the number of total ankle arthroplasties (TAAs) performed increases each year, so too does the number requiring revision. Previous studies have demonstrated that revision of these implants is a challenging procedure, with a high incidence of complication and reoperation. Conversion to a tibiotalocalcaneal (TTC) arthrodesis or an ankle arthrodesis is often favored in patients with profound bone loss. The factors affecting a patient’s likelihood of undergoing a reoperation after TAA revision to an arthrodesis are not well described. This study aims to examine the factors influencing the likelihood of reoperation following TAA revision to an arthrodesis. Methods: This is a single-institution, retrospective cohort study of patients who underwent TAA revision to an arthrodesis between 2002 and 2023. Cases were identified using CPT codes (27703, 27704) and screened to include only those who underwent revision of a total ankle arthroplasty to an arthrodesis. Fifty unique ankles met these criteria. The average duration of follow-up was 52.9 months. The primary outcome was reoperation, defined as any unplanned procedure performed on the ipsilateral ankle due to implant complications. Collected variables included patient age, duration since primary TAA, type of structural allograft, and fixation construct, among others. Statistical analyses were performed in SPSS using binary logistic regression or chi-square, as appropriate. Results: Search and screening yielded 50 unique total ankle arthroplasty revisions to an arthrodesis. Of these, 20 required reoperations, most often for the removal of symptomatic implants (10). Patient age had a weak positive effect on the likelihood of reoperation (B=0.035, OR=1.036), however this effect was not statistically significant (p=0.206). Increased time to revision had a weak negative effect on the likelihood of reoperation (B = -0.090, OR = 0.914), but this association was also not statistically significant (p = 0.110). The relative risk of reoperation when using a femoral head graft compared to a tricortical bone graft was 1.730 (95% CI: 0.607-4.929). The relative risk of reoperation when using only intramedullary fixation compared to any extramedullary fixation was 0.927 (95% CI: 0.352-2.436). Conclusion: Outcomes following the revision of TAAs to an arthrodesis remain difficult to predict. Our findings are limited by low power and larger cohorts may yield significant findings. Furthermore, our cohort included only one TTC arthrodesis using a cage construct, preventing us from drawing conclusions about this strategy. Future directions of study include examining the influence of radiographic parameters. Given the high rates of reoperation observed, patients should be counseled that revision to an arthrodesis does not guarantee the avoidance of future procedures.