Abstract
Background:
Distal tibiofibular synostosis is a known but poorly described complication of ankle fractures. The objective of this study was to evaluate the relationship between ankle fracture fixation method and other risk factors in the development of synostosis in posttraumatic operative ankle fractures. Outcomes of patients with no synostosis, incomplete bony bridging, or complete synostosis also were compared.
Methods:
All operative ankle fractures from a single surgeon’s prospectively collected clinical database were evaluated for the presence of distal tibiofibular synostosis not earlier than 3 months postoperatively. All fractures were treated in a fracture-specific and ligament-specific fashion. Syndesmotic screws, when used, were routinely removed no sooner than 4 months after operation. Patient demographic and medical history data as well as injury and fixation profile were evaluated. Incidence of complications was also reviewed. Foot and Ankle Outcome Score (FAOS) and range of motion (ROM) were compared between the groups.
Results:
Of the 564 ankle fractures included in the study, 91 demonstrated complete synostosis and 46 demonstrated incomplete bony bridging. Multivariate analysis revealed male sex (odds ratio [OR] = 2.82, P < .001), syndesmotic screw fixation (OR = 2.46, P < .001), and tibiotalar dislocation (OR = 1.74, P = .032) to remain significant independent risk factors for the development of incomplete bony bridging or complete synostosis while adjusting for confounding risk factors. Ankles with synostosis also demonstrated significant reduction in dorsiflexion (P = .017), plantarflexion (P = .024), and inversion (P < .001), though patient-reported outcome measures were equivalent with patients without synostosis.
Conclusion:
Syndesmotic screw fixation, male sex, and tibiotalar dislocation were significant risk factors in the formation of postoperative distal tibiofibular synostosis, though patient-reported outcome measures seemed equivalent among patients with and without synostosis.
Level of Evidence:
Level III, case control study of prospectively collected registry data.
Introduction
Ankle fractures are common orthopaedic injuries with an incidence of 184 fractures per 100 000 persons per year. 14 Distal tibiofibular synostosis is a known but poorly described complication of ankle fractures. Incidence in all ankle fractures is reported as 3% to 7%, with up to 12% of Weber C fractures displaying synostosis.1,2,7 Most ankles that develop synostosis do so by 3 months postoperatively. 1 Clinically, patients may be completely asymptomatic or may complain of pain about the ankle with weightbearing and ankle push off.4,15 The literature is scant and lacks consensus concerning the correlation of synostosis with the use of syndesmotic screws for fixation. It has yet to be delineated whether synostosis results solely from syndesmotic injury, use of syndesmotic screws for fixation, or a complex interplay of the two.
The objective of this study was to evaluate the relationship of ankle fracture fixation method (syndesmotic screw fixation vs deltoid ligament repair vs open reduction and internal fixation of posterior malleolar fractures) with the presence of synostosis in posttraumatic operative ankle fractures and compare outcomes. We hypothesized that repair with syndesmotic screw fixation would be a risk factor for the development of distal tibiofibular synostosis.
Methods
This study was a case-control and outcomes study. Institutional review board approval was obtained, and we reviewed a prospectively created clinical database of 748 surgically treated ankle fractures by the senior author from January 2003 to April 2013. We reviewed the latest ankle radiographs for distal tibiofibular synostosis, classifying ankles as (1) no synostosis, (2) incomplete bony bridging, or (3) complete synostosis (Figures 1-3). Ankles were classified as having complete synostosis if continuity of a complete bony bridge was observed on all 3 standard ankle series radiographs, anteroposterior (AP), mortise, and lateral ankle radiographs, obtained at the latest postoperative visit. Incomplete bony bridging was noted if full bony continuity was not observed on any 1 of the 3 standard ankle series radiographs. Ankles were classified as having no synostosis if absolutely no ectopic bone formation was present within the distal tibiofibular interval.
Plain radiograph of a SER IV fracture displaying no synostosis at 11 months postoperatively.
Plain radiograph of a SER IV fracture displaying incomplete bony bridging at 3 months postoperatively.
Plain radiograph of a variant ankle fracture displaying complete synostosis at 9 months postoperatively.
Inclusion criteria included age ≥ 18 years and a minimum 3 months of radiographic follow-up. Ten patients were excluded due to age and 174 patients were excluded due to inadequate radiographic follow-up. Of the 748 surgically treated ankle fractures, 564 met inclusion criteria. Of the 564 ankle fractures included in the study, 43% (244 of 564) occurred in men. Mean age at time of surgery was 51 years (range, 18-91). Mean duration of follow-up for functional outcomes was 25 months (range, 12-110). Ninety-one ankles demonstrated complete synostosis, 46 ankles demonstrated incomplete bony bridging, and 427 ankles demonstrated no synostosis.
All fractures were treated in a fracture-specific and ligament-specific manner with the patient positioned prone on the operating room table. A posterolateral approach was typically performed via a longitudinal incision approximately 10 cm in length between the posterior border of the fibula and the lateral border of the Achilles tendon, curved anteriorly at the distal aspect of the fibula. Scalpel dissection was taken through fascia, and the interval between the flexor hallucis longus (FHL) and peroneus longus and brevis was identified. Dissection of the intramuscular fascial plane continued down to the fibula, and the peroneus longus and brevis were retracted anteriorly with the FHL retracted posteriorly. Subperiosteal elevation of the musculature was then performed to provide exposure to the posterior tibial and fibular fracture apices. Fibular fractures were treated with plate fixation with either syndesmotic screw fixation or deltoid ligament repair with suture anchor whenever the syndesmosis was found to be unstable. Reduction and rigid fixation of all posterior injuries were performed to improve syndesmotic stability with either a small fragment antiglide plate for posterior malleolar fractures, a 3.5 mm cortical screw with washer for isolated posteroinferior tibiofibular ligament (PITFL) tears, or 2 fragment antiglide plates for posterolateral and posteromedial fragments of hyperplantarflexion variant fractures. Postoperatively, all patients were immobilized in a lower leg splint and placed on a nonweightbearing protocol. At the 2-week postoperative visit, patients were advanced to a boot, and ankle range of motion (ROM) exercises were initiated. At the 6-week postoperative visit, patients were typically progressed to weightbearing as tolerated. Syndesmotic screws, when present, were routinely removed no sooner than 4 months after operation. 11
Patient demographic data included sex and age, and medical history included diabetes (DM), hypertension (HTN), hyperlipidemia (HLD), history of tobacco use, peripheral vascular disease (PVD), and Body Mass Index (BMI). Postoperative infection and wound complications were also noted. Injury profile included fracture side (left vs right), Lauge-Hansen classification, Weber classification, geriatric fracture (age 65 or older), and tibiotalar dislocation. Injury profile was determined via a combination of preoperative radiographs, magnetic resonance imaging (MRI) and/or computed tomography (CT) evaluation, and intraoperative assessment. Fixation profile included syndesmotic screw fixation, deltoid ligament repair, and/or posterior malleolus osteosynthesis.
For the outcome portion of the study, primary functional outcome, as defined by the Foot and Ankle Outcome Score (FAOS), was evaluated. This validated patient-centered questionnaire measured function subcategorized as symptoms, pain, activities of daily living, sports/recreation, and quality of life.5,13 Secondary outcome included range of motion. Inclusion criteria for this portion of the study was a minimum of 1 year of clinical follow-up.
Stastical Methods
Statistical analyses were first performed to determine risk factors associated with the degree of synostosis (none, incomplete, or complete). Analyses were also performed to determine differences in postoperative ROM, frequency of complications, and differences in patient-reported outcome measures in patients with at least 1 year of follow-up. Associations between categorical variables and degree of synostosis were determined with chi-square or Fisher’s exact tests. Post hoc analyses for these variables were performed with the Bonferroni-corrected, pairwise comparisons. Relationships between continuous variables and degree of synostosis were evaluated via single-factor analysis of variance (ANOVA) with Tukey post hoc tests. If the continuous variable was non-normally distributed, Kruskal Wallis tests were utilized with Bonferroni-corrected, post hoc Mann-Whitney U tests.
Stepwise multivariate ordinal logistic regression was then performed to determine significant independent risk factors while adjusting for confounding risk factors. Risk factors considered were those with a P < .25 from the univariate analyses. The level of significance for all tests was P < .05. Descriptive statistics are shown as odds ratios (OR) with 95% confidence intervals (CI) for categorical variables and means ± standard deviations (SD) for continuous variables. Statistical analyses were performed with SAS version 9.3 (Cary, North Carolina).
Results
Incidence of complete synostosis was 16.1% (91 of 564). Incidence of incomplete bony bridging was 8.2% (46 of 564). Table 1 summarizes the demographic characteristics, injury profile, and fixation profile data. Males were more likely than females to develop incomplete bony bridging or complete synostosis (OR = 2.19; 95% CI, 1.18-4.05 and OR = 3.11; 95% CI, 1.94-4.99, respectively; P < .001). There was no significance of age at time of surgery or fracture side in the development of synostosis.
Patient demographic characteristics, injury profile, and fixation profile summary.
Pronation external rotation IV (PER IV) fractures were more at risk for developing incomplete bony bridging and complete synostosis (OR = 3.65; 95% CI, 1.84-7.23 and OR = 2.70; 95% CI, 1.56-4.69, respectively; P < .001). Supination external rotation IV (SER IV) and variant fractures did not show any statistically significant difference. No patients with Supination adduction (SA) or SER III fractures developed complete synostosis; however, the small numbers of patients sustaining these and other less common Lauge-Hansen fracture patterns precluded detection of significant differences. Fractures categorized as Weber C were also more likely to develop incomplete bony bridging or complete synostosis (OR = 2.32; 95% CI, 0.51-10.65 and OR = 2.56; 95% CI, 0.72-9.06, respectively; P = .023). Geriatric fractures did not show a significantly increased risk of developing synostosis. Tibiotalar dislocation was associated with increased odds of incomplete bony bridging and complete synostosis (OR = 1.10; 95% CI, 0.54-2.24 and OR = 1.99; 95% CI, 1.23-3.23, respectively; P = .019), and the use of syndesmotic screws was associated with increased odds of incomplete bony bridging and complete synostosis (OR = 2.07; 95% CI, 1.11-3.86 and OR = 3.16; 95% CI, 1.94-5.15, respectively; P < .001). There were no significantly increased odds of developing synostosis among the other fixation methods.
Table 2 summarizes the patient medical history and postoperative complication data. Patients with elevated BMI were more likely to demonstrate incomplete bony bridging and complete synostosis versus no synostosis (BMI 30.5 and 29.1, respectively, vs 28.0; P = .040). Patients with postoperative infections were more likely to develop incomplete bony bridging and complete synostosis (OR = 2.44; 95% CI, 0.78-7.63 and OR = 2.85; 95% CI, 1.26-6.42, respectively; P = .020). There was no significantly increased risk of developing synostosis in patients with DM, HTN, HLD, history of tobacco use, PVD, or postoperative wound complications.
Patient medical history and complications summary.
Abbreviations: DM, diabetes; HTN, hypertension; HLD, hyperlipidemia; PVD, peripheral vascular disease; BMI, Body Mass Index.
Table 3 summarizes multiple stepwise logistic regression analysis. Male sex (OR = 2.82; 95% CI, 1.77-4.52; P < .001), syndesmotic screw fixation (OR = 2.46; 95% CI, 1.53-3.95; P < .001), and tibiotalar dislocation (OR = 1.74; 95% CI, 1.05-2.86; P = .032) all remained significant independent risk factors for the development of incomplete bony bridging or complete synostosis after adjusting for confounding risk factors.
Multivariate analysis of significant risk factors, adjusted for confounding risk factors.
Tables 4 and 5 summarize primary and secondary outcome data, respectively. There were no significant differences among the no synostosis, incomplete bony bridging, or complete synostosis groups concerning FAOS. Patients with complete synostosis demonstrated reduced dorsiflexion (16 degrees vs 18 degrees; P = .017), plantarflexion (44 degrees vs 47 degrees; P = .024), and inversion (24 degrees vs 30 degrees; P < .001) compared with patients without synostosis. There was no significant difference in eversion among the 3 groups.
Foot and Ankle Outcomes Score (FAOS), presented as mean (SD).
Abbreviations: ADL, activities of daily living; QOL, quality of life.
Range of motion (ROM), presented as mean degrees (SD).
Discussion
The objective of this study was to ascertain risk factors for the development of synostosis and compare functional outcomes between patients with no synostosis, incomplete bony bridging, and complete synostosis following operative treatment of ankle fractures. The null hypothesis that syndesmotic screw fixation was not a risk factor for the development of distal tibiofibular synostosis was rejected and our experimental hypothesis accepted.
Reported incidence of synostosis in operative ankle fractures is 3% to 7%.1,2,7 Our study found an incidence of complete synostosis of 16.1%. Some degree of postoperative bony bridging was present in 24.3% (137 of 564) of the ankle fractures. Including the 184 patients that were excluded due to age and lack of adequate radiographic follow-up, the incidence of synostosis was 12.2%. The discrepancy in rates of synostosis is likely multifactorial but may be affected by the number of excluded patients and an increased frequency of high energy trauma.
Three months of radiographic follow-up was chosen as the cutoff for inclusion in the review for synostosis risk factors. Albers et al noted that 86.7% of their patients that developed synostosis did so within 3 months with all 15 ankle fractures displaying synostosis by 1 year postoperatively. 1 Harborne and Lennox describe the formation of distal tibiofibular synostosis within 4 months of injury in their case report of a 21-year-old male who developed synostosis after a direct blow. 6 Hou et al noted that distal tibiofibular synostosis was present in their cohort of 8 patients within 12 weeks of operation. 7
To the best of our knowledge, male sex and tibiotalar dislocation as risk factors for the development of distal tibiofibular synostosis was not previously reported. In our study, multivariate analysis identified males as having almost 3 times the odds of developing synostosis compared to females. This relationship may be due to the increased occurrence of high energy trauma in males. Additionally, high energy trauma may more frequently result in tibiotalar dislocation and is likely a common underlying etiology. This conclusion is further supported by our finding that PER IV and Weber C fractures, often associated with high energy mechanisms, demonstrate significant association with the development of tibiofibular synostosis.
Our finding that syndesmotic screw fixation was a significant risk factor in the development of synostosis is contrary to some reports in the literature. Albers et al found no relationship between syndesmotic screw fixation versus suture reconstruction of the syndesmosis in the development of synostosis. 1 Böstman noted development of synostosis in 3 of 80 patients with malleolar fractures without tibiofibular diastasis treated with osteosynthesis. Though these patients did not undergo syndesmotic screw fixation, absorbable polyglycolide rods were used for fixation and may have contributed to the development of synostosis. 3 Although a complex interplay between syndesmotic injury, syndesmotic screw fixation, and synostosis likely exists, the comparable extent of injury among ankles in our study minimized this potentially confounding effect. Additionally, management of syndesmotic injury has changed at our institution. Previously, transsyndesmotic screws were employed for stabilization, however, over time posterior malleolar osteosynthesis or PITFL repair has become our treatment of choice for most syndesmotic injuries. Thus, our evaluation was well balanced when comparing cases treated with and without syndesmotic screw fixation.
Similar to previous studies, we found no difference in terms of patient-reported outcome measures among patients with and without distal tibiofibular synostosis. Albers et al employed a modified ankle score derived from a functional ankle score described by Philips et al. 12 They reported a mean ankle score of 91 out of a possible 100 (range, 71-100) in patients with synostosis versus mean ankle score of 92 (range, 68-100) in all 230 of their ankle fracture patients. 1 Hou et al employed a functional ankle score described by Mazur et al with a maximum score of 100 and scores from 80 to 90 considered excellent. 9 They found a mean score of 88.1 in their review of 8 patients with postoperative synostosis. 7
There are conflicting reports in the literature regarding the effect of synostosis on ankle ROM. McMaster and Scranton noted that in normal fibular motion, the fibula is pulled distally, tightening the interosseous membrane, and increases stability during weightbearing. They argue that a synostosis alters this normal motion by preventing fibular descent thus reducing dorsiflexion. 10 Hou et al reported an 8.26 degree reduction in dorsiflexion compared to the contralateral ankle in their review of 8 patients with postoperative synostosis. 7 However, van den Bekerom et al argued that tibiofibular synostosis is unlikely to disrupt normal ankle motion because syndesmotic screw fixation of an anatomically reduced ankle fracture will not restrict dorsiflexion. 2 Furthermore, Albers et al’s review of 230 ankles showed no difference in ROM between ankles with or without synostosis. 1 Our study found that patients with distal tibiofibular synostosis demonstrated an average 2 degree reduction in dorsiflexion as compared with ankles without synostosis. We also found a restriction of plantarflexion and inversion in patients with synostosis.
Clinically, patients with distal tibiofibular synostosis typically have little to no complaints. 4 Rarely does distal tibiofibular synostosis require operative treatment. However, surgical excision of a synostosis may benefit symptomatic young and active sportsmen and sportswomen.6,8,15 The goal of operative treatment is to completely excise the synostosis and return physiologic external rotation and distal translation to the fibula. The mechanism of synostosis formation is poorly understood. McMaster and Scranton suggested synostosis occurs secondary to ossification of hematoma resulting in fusion of the tibia and fibula. 10 Hou et al proposed 2 possible etiologies: (1) hematoma absorption and ossification following laceration of the interosseous membrane between distal tibia and fibula and (2) the damage of the interosseous membrane caused by Kirschner wires or screws. They also suggested that periosteal stripping during the operation may play a role. 7
There are limitations to the current study. As a risk factor review study, all treatments were at the discretion of the operating surgeon, the senior author, and thus were not under experimental control. Furthermore, the senior author was the single operating surgeon at the same institution who treated all patients, which perhaps may reduce the generalizability of these results. However, this also demonstrates consistency in evaluating, treating, and following the patients clinically and is a major strength of the study. Nearly all operations were performed via the posterolateral approach. Thus, too few operations were performed via alternative approaches to compare the potential role of surgical approach in the development of synostosis.
Our study produced several interesting findings. Only syndesmotic screw fixation, not deltoid ligament repair or posterior malleolus repair, demonstrated a significant association with the development of tibiofibular synostosis. However, 27 ankle fractures, all stage IV injuries, did not undergo syndesmotic screw fixation yet developed synostosis. We conclude that while syndesmotic screw fixation obviously plays a significant role in the development of synostosis, high energy trauma (as evidenced by the significant association of male sex, tibiotalar dislocation, PER IV fracture designation, and Weber C fracture designation) presumably plays some minor role. Further investigation is needed to completely establish the complex interplay between syndesmotic injury, syndesmotic screw fixation, and synostosis.
Synostosis seems to occur in a continuum of ectopic bone formation in the distal tibiofibular interval from incomplete bony bridging to full bony bridging with complete bony continuity. To the best of our knowledge, this current study was the first to identify male sex, syndesmotic screw fixation, and tibiotalar dislocation as significant independent risk factors in the formation of distal tibiofibular synostosis. Patients with PER IV fractures, Weber C fractures, postoperative infections, or elevated BMI were all at increased risk for developing synostosis. We also noted reduction of dorsiflexion, plantarflexion, and inversion in ankles with synostosis, though patient-reported outcome measures were equivalent to patients without synostosis.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.