Revision of ankle arthrodesis.

From 1989 to 1996, we treated 18 cases (10 males, eight females; average age 48.2 years) of failed ankle arthrodesis by revision of ankle arthrodesis and followed their progress for at least two years. The average time interval between original surgery and revision was 17.3 months. Revisions were needed due to infection in one case, nonunion in 10 cases, and malalignment in seven cases. The salvage operations included debridement in the infected case, refreshed pseudoarthrosis in nonunion cases, and corrective osteotomy in malalignment cases. Sixteen cases were fixed by crossed screws with internal compression, one infected case was fixed by an external fixator, and one case with bone loss was fixed with buttress plate. The average follow-up period was 40.4 months. There was one nonunion and two delayed unions, with an ultimate fusion rate of 94%. The average AOFAS ankle-hindfoot score was 70.9 at final follow up. There was one excellent result (5.6%), five good results (27.8%), 11 fair results (61%), and one poor result (5.6%), and the overall results were poorer compared with our series of primary arthrodesis. The time to fusion also took longer in the revision cases (average 2.7 months in primary cases and 4.8 months in revision cases). Fusion techniques that ensure solid union in a functional position are essential. If an ankle arthrodesis fails, however, revision is a salvage procedure that can achieve an acceptable result.