What Are the Indications for Implant Revision in Three-component Total Ankle Arthroplasty?

BACKGROUND: Given the increasing usage of total ankle arthroplasty (TAA), a better understanding of the reasons leading to implant revision and the factors that might influence those indications is necessary to identify at-risk patients. QUESTION/PURPOSES: Using a single-design three-component ankle prosthesis, we asked: (1) What is the cumulative incidence of implant revision at 5 and 10 years? (2) What are the indications for implant revision in our population? (3) What factors are associated with an increased likelihood of implant revision during the time frame in question? METHODS: Between 2003 and 2017, primary TAA using a single-design three-component ankle implant was performed by or under the supervision of the implant designer in 1006 patients (1074 ankles) aged between 17 and 88 years to treat end-stage ankle osteoarthritis. No other TAA systems were used during the study period at the investigators' institution. In 68 patients with bilateral surgery, only the first TAA was considered. Of the patients treated with the study implant, 2% (16 of 1006) were lost to follow-up 5 to 14 years after TAA and were not known to have died or undergone revision, and 5% (55 of 1006) were deceased due to reasons unrelated to the procedure, leaving 935 patients for evaluation in this retrospective study. The mean (range) follow-up for the included patients was 8.8 ± 4.2 (0.2 to 16.8) years. Implant revision was performed 0.5 to 13.2 years after TAA in 12% (121 of 1006) of our patients. Survivorship free from revision was calculated using cumulative incidence (competing risks) survivorship, with death as a competing risk. The reason for each revision was classified into one of six categories according to a modified version of a previously published protocol: aseptic loosening, cyst formation, instability, deep infection, technical error, and pain without another cause. Two foot and ankle surgeons reviewed the records of all patients who underwent implant revision and assigned each patient's reason for revision to one of the six categories. The decision for assigning each patient to one of the six categories was made based on a consensus agreement. A subgroup classification of preoperative ankle alignment (neutral, mild, and major deformity) and variables of age, sex, BMI, etiology of ankle osteoarthritis, and number of preoperative and intraoperative hindfoot or midfoot procedures were used in a multinomial logistic regression and Cox regression analysis to estimate their association with reason for revision and implant survival until revision. RESULTS: The cumulative incidence of implant revision at the mean (range) follow-up time of 8.8 ± 4.2 years (0.2 to 16.8) was 9.8% (95% confidence interval 7.7% to 11.8%). Five and 10 years after TAA, cumulative incidence was 4.8% (95% CI 3.4 to 6.1) and 12.1% (95% CI 9.7% to 14.5%), respectively. The most common reason for revision was instability (34% [41 of 121]), followed by aseptic loosening of one or more metallic components (28% [34 of 121]), pain without another cause (12% [14 of 121]), cyst formation (10% [12 of 121]), deep infection (9% [11 of 121]), and technical error (7% [9 of 121]). Ankles with a major hindfoot deformity before TAA were more likely to undergo revision than ankles with a minor deformity (hazard ratio 1.9 [95% CI 1.2 to 3.0]; p = 0.007) or neutral alignment (HR 2.5 [95% CI 1.5 to 4.4]; p = 0.001). A preoperative hindfoot valgus deformity increased revision probability compared with a varus deformity (HR 2.1 [95% CI 1.4 to 3.4]; p = 0.001). CONCLUSION: Instability was a more common reason for implant revision after TAA with this three-component design than previously reported. All causes inducing either a varus or valgus hindfoot deformity must be meticulously addressed during TAA to prevent revision of this implant. Future studies from surgeons/institutions not involved in this implant design are needed to confirm these findings and to further investigate why a substantial number of patients had pain of unknown cause prompting revision. LEVEL OF EVIDENCE: Level III, therapeutic study.