Risk Factors for Aseptic Revision of Operatively Treated Ankle Fractures

Background: Early revision rates within 12 months after ankle fracture open reduction internal fixation (AF-ORIF) are fairly low; however, they remain relevant given the volume of ankle fractures occurring annually. Understanding these rates is complex because reoperation due to technical or mechanical complications is typically reported alongside soft tissue–related problems such as symptomatic hardware, wound dehiscence, or infection. There are limited data identifying risk factors specifically for revision of ankle fracture fixation in the absence of soft tissue complications. Understanding variables that predispose to aseptic technical and mechanical failure without this confounder may provide insight and improve patient care. Methods: A retrospective cohort study was performed at 2 large academic medical centers. Research Patient Data Registry (RPDR) data available from 2002 to 2019 were used to identify patients who underwent aseptic revision of AF-ORIF within 12 months (n = 33). A control group (n = 100) was selected by identifying sequential patients who underwent AF-ORIF not requiring revision within 12 months. Multiple factors were recorded for all patients in both cohorts. Each fracture was also evaluated according to the Ankle Reduction Classification System (ARCS) of Chien et al,8 which categorizes biplanar talar displacement in relation to a central tibial plumb line into 1 of 3 grades: A (0-2 mm), B (3-10 mm), and C (>10 mm). Adapted from its original purpose of grading reduction quality, we applied ARCS to pre-reduction radiographs to assess initial fracture displacement. All variables collected were compared in univariate analysis. Variables that achieved significance in univariate comparisons were included as candidates for multivariable analysis. Results: Final multivariable logistic regression modeling demonstrated the following factors to independently predict the need for aseptic revision surgery: documented falls in the early postoperative period (aOR, 298; 95% CI, 15.4, 5759; P < .001), movement-altering disorders (aOR, 81.7; 95% CI, 4.12, 1620; P = .004), a nonanatomic mortise (medial clear space [MCS] > superior clear space [SCS]) on immediate postoperative imaging (aOR, 38.4; 95% CI, 5.53, 267; P < .001), initial coronal plane tibiotalar displacement >10 mm and sagittal plane tibiotalar dislocation (ARCS-C) (aOR vs ARCS-A, 25.8; 95% CI, 2.81, 237; P = .004), substance abuse (aOR, 15.7; 95% CI, 2.66, 92.8; P = .002), and polytraum