Introduction

The ankle syndesmosis is a fibrous articulation joining the distal tibia and fibula that is stabilized by three ligaments including the anterior–inferior tibiofibular ligament (AITFL), posterior–inferior tibiofibular ligament (PITFL), and interosseous tibiofibular ligament (ITFL) [1, 10, 44]. The space between the tibia and fibula form the synovial recess which contains the direct articulating cartilage surfaces of the tibia and fibula, described previously as the tibiofibular contact zone [1, 10, 44]. Together, these elements comprise the primary structures of the ankle syndesmosis.

Sprains of the ankle syndesmosis, commonly called high ankle sprains, account for as much as 25 % of all ankle sprains in athletic patient populations [19]. Most isolated sprains of the syndesmosis may be treated non-operatively; however, syndesmosis injuries can often result in prolonged periods of pain and functional limitations [7, 13, 17, 39]. Furthermore, patients with tibial or fibular fractures and concomitant syndesmotic sprains, or patients with grade III acute or chronic syndesmotic instability often require surgical treatment ranging from proximally placed indirect fixation (screws or suture-button constructs) to allograft reconstruction. Previous clinical outcome studies have correlated anatomic reduction of the syndesmosis with improved clinical outcomes following surgery [30, 37]. Despite this caveat, malreduction is common and current methods to confirm an anatomic reduction are not always accurate whether that be through routine radiographs, fluoroscopy, or stress radiographs [5, 6, 8, 11, 12, 18, 23, 31].

Anatomic reduction is predicated on accurate identification of native syndesmosis anatomy. Recent cadaveric studies have outlined qualitative and quantitative descriptions of syndesmosis anatomy for use during anatomic-based surgical repair and reconstruction procedures [1, 10, 44]. However, radiographic guidelines detailing the anatomic attachments of the syndesmotic ligaments and location of the syndesmotic articular cartilage surfaces are currently lacking. Radiographic guidelines would augment current diagnostic approaches, improve preoperative planning, assist with intraoperative identification of native anatomy, and facilitate objective postoperative assessment of anatomic-based reduction, repair, and reconstruction techniques. Radiographic data describing the anatomic locations of the structures of the syndesmosis may be particularly useful in revision cases or those with concomitant injury where other anatomic landmarks and navigation techniques may be more difficult to interpret.

Therefore, the purpose of this study was to establish qualitative and quantitative radiographic guidelines for identifying the tibial and fibular attachments of the three syndesmotic ligaments and the articulating surfaces of the syndesmosis using standard ankle radiographic views. It was hypothesized that these sites could be reproducibly defined in relation to osseous landmarks and superimposed radiographic axes.

Materials and methods

Twelve non-paired, fresh-frozen cadaveric specimens (mean age 56, range 38–82 years; 4 females and 8 males; 8 left and 4 right) with no history of ankle injury, surgery, osteoarthritis, or significant anatomic abnormalities were used in this study. This sample size was based on similar previously published research [16]. De-identified cadaveric specimens are exempt from Institutional Review Board (IRB) review at our institution; therefore, IRB approval was not required for this study. The relative anatomic positions of the tibia and fibula were preserved using rigid screw fixation placed 10 and 15 cm proximal to the tibiotalar joint line. Soft tissue dissections were subsequently performed to identify the origin and insertion sites of the AITFL, PITFL, and ITFL in accordance with previous anatomic literature [1, 10, 44]. The three syndesmotic ligaments were sequentially transected at their midsubstance. Tibial and fibular ligamentous remnants were then used to identify the tibial and fibular attachment sites or footprints. The centres of the AITFL and PITFL footprints were identified and marked by shallowly embedding a 2-mm stainless steel sphere (diameter: 2.0 ± 0.0025 mm, sphericity: 0.0006 mm, Small Parts, Inc., Logansport, IN) similar to a previously described technique [16]. Due to the multifascicular nature and/or broad sites of attachment of the AITFL and PITFL, additional 2-mm stainless steel spheres were placed at the superior and inferior bands/margins of each origin and insertion. For the AITFL, this included 2-mm spheres placed in the proximal and distal accessory band(s) (Bassett’s Ligament) [2, 44]. For the PITFL, this included the proximal and distal borders of the superficial PITFL in addition to tibial and fibular footprint centres of the deep PITFL fibres (inferior transverse tibiofibular ligament). For the ITFL, the proximal and distal extents of the tibial and fibular fibre attachments were marked using the same technique. Stainless steel spheres were also embedded in the centre of the tibial and fibular cartilage-covered articulating facets, described previously as the tibiofibular contact zone [1, 10, 44]. To ensure that individual spheres could be distinguished in the event of overlap on mortise and lateral radiographs, spheres were placed sequentially from anterior to posterior in each ligament/structure with sequential radiographs obtained at these intervals. Sequential mortise and lateral radiographs were superimposed and compared to accurately identify the individual metallic spheres representing each respective component (AITFL, ITFL, PITFL, Tibiofibular contact zone) of the ankle syndesmosis.

Data collection

Standard lateral and mortise radiographs of each specimen were obtained using a fluoroscopic mini-C-arm (Hologic, Inc., Bedford, MA). Images were obtained under live fluoroscopy to obtain true lateral and mortise views [27]. Lateral view radiographs were defined by an X-ray beam coincident with the intermalleolar axis and superimposition of the medial and lateral profiles of the talar dome. Mortise view radiographs were defined by an X-ray beam perpendicular to the intermalleolar axis and clear visualization of the talofibular joint space. A 25.4-mm-diameter radiopaque stainless steel sphere (diameter: 25.4 ± 0.00254 mm, sphericity: 0.00061 mm, Small Parts, Inc., Logansport, IN) placed at the level of the ankle joint was utilized in all radiographs for measurement calibration and to adjust for any differences in magnification caused by variation in specimen distance from the X-ray source [16].

Radiographic images were then imported into a picture archiving and communication system (PACS) for measurements (eFilm Workstation® 3.4, Merge Healthcare Inc., Chicago, IL). Radiographic landmarks were selected, and measurements were taken under the direction of a foot and ankle fellowship trained orthopaedic surgeon and the senior author (TOC). These radiographic landmarks are depicted in Fig. 1. The medial–lateral axis for mortise views and anterior–posterior axis for lateral views were defined by a superimposed reference line parallel to and at the level of the tibial plafond (Fig. 2). The superior-inferior axis was defined by a superimposed reference line perpendicular to the tibial plafond reference line and coincident with the long axis of the tibia [16].

Fig. 1
figure 1

Representative A lateral and B mortise radiographic views with labelled reference landmarks used to quantitatively characterize the locations of individual syndesmotic structures. a Anterior tibial plafond; b tibial plafond; c posterior tibial plafond; d anterior fibular tubercle; e inferior tip of the lateral malleolus; f medial corner of the tibial plafond; g lateral corner of the tibial plafond; h most lateral tibial point

Full size image
Fig. 2
figure 2

Representative a lateral and b mortise radiographic views demonstrating the axes used for radiographic measurements

Full size image

Statistical analyses

Measurements were taken by two independent observers with varying levels of medical training to calculate interobserver reliability (BTW, KAJ). Measurements included the mean absolute distance in addition to the mean superior–inferior component, and the mean anterior–posterior (lateral view) or medial-lateral (mortise view) component of each distance. Agreement between reviewers and across trials was assessed via 2-way mixed, random measure intraclass correlation coefficients (ICCs) for each ligament/structure and radiographic view [38]. Statistical analyses were performed using SPSS Statistics, version 20 (SPSS Inc, an IBM Company). For calculation of the intraobserver ICCs, the primary reviewer (BTW) performed measurements twice separated by a minimum interval of 2 weeks to reduce the potential for recall bias.

Results

Lateral radiographic view

Select distances from each syndesmotic structure to individual radiographic landmarks on the lateral view are reported in Tables 1, 2 and 3 as means and standard deviations and visually represented in Figs. 1A, 3a, 4a and 5a. Both interobserver and intraobserver ICCs demonstrated excellent agreement between raters and reproducibility across trials for all structures of the syndesmosis on lateral radiographic views (Tables 4, 5).

Table 1 Radiographic measurements of the anterior–inferior tibiofibular ligament, lateral view
Full size table
Table 2 Radiographic measurements of the posterior–inferior tibiofibular ligament, lateral view
Full size table
Table 3 Radiographic measurements of the interosseous tibiofibular ligament and tibiofibular contact zone, lateral view
Full size table
Fig. 3
figure 3

Representative a lateral and b mortise radiographic views demonstrating the attachment sites of the anterior–inferior tibiofibular ligament (AITFL), including the tibial and fibular attachment centres of the proximal accessory bands (A T1/A F1), primary bands (A T2/A F2), and the distal accessory (Bassett’s ligament) band (A T3/A F3)

Full size image
Fig. 4
figure 4

Representative a lateral and b mortise radiographic views demonstrating the posterior–inferior tibiofibular ligament (PITFL) attachment sites including the superficial and deep components. The proximal (P T1/P F1) and distal (P T3/P F3) margins of the superficial PITFL are indicated in addition to its tibial and fibular footprint centres (P T2/P F2). The centres of the tibial and fibular deep attachments are also labelled (P TD/P FD)

Full size image
Fig. 5
figure 5

Representative a lateral and b mortise radiographic views demonstrating the proximal (I T1/I F1) and distal (I T2/I F2) extents of the tibial and fibular attachments of the interosseous tibiofibular ligament (ITFL) in addition to the articular cartilage facets (CZT/CZF) of the tibiofibular contact zone

Full size image
Table 4 Interobserver  reliability
Full size table
Table 5 Intraobserver reproducibility
Full size table

On the lateral view (Fig. 3a), the AITFL tibial attachment was superior and slightly posterior to the anterior corner of the tibial plafond, while the AITFL fibular footprint centre was superior and posterior to the anterior-most point of the anterior fibular tubercle. The superficial PITFL (Fig. 4a) footprint centre was superior to the posterior corner of the tibial plafond, while the deep PITFL attached further distally and anteriorly. The ITFL (Fig. 5a) had a broad tibial attachment, extending from 45.9 ± 7.9 mm proximal to the joint line to 12.4 ± 3.4 mm proximal to the joint line as measured in line with the long axis of the tibia. Distal to the inferior margin of the ITFL, a synovial-lined joint space, which contained areas of tibial and fibular articulating cartilage (Fig. 5a), termed the syndesmotic tibiofibular contact zone, were found in all specimens. The centre of the tibial articulating cartilage was posterior and superior to the anterior corner of the tibial plafond.

Mortise radiographic view

Relevant distances from each syndesmotic structure to select radiographic landmarks on the mortise view are listed in Tables 6, 7 and 8 as means and standard deviations and can be visualized in Figs. 1B, 3b, 4b and 5b. Interobserver and intraobserver ICCs both demonstrated excellent agreement between raters and reproducibility across trials for all structures of the syndesmosis on mortise views (Tables 4, 5).

Table 6 Radiographic measurements of the anterior–inferior tibiofibular ligament, mortise view
Full size table
Table 7 Radiographic measurements of the posterior–inferior tibiofibular ligament, mortise view
Full size table
Table 8 Radiographic measurements of the interosseous tibiofibular ligament and tibiofibular contact zone, mortise view
Full size table

On the mortise view, the AITFL (Fig. 3b) coursed distally and laterally from its tibial origin, which was lateral and superior to the lateral corner of the tibial plafond. The PITFL (Fig. 4b) coursed distally and laterally from its tibial origin to fibular insertion. The centre of the superficial PITFL tibial footprint was medial and superior to the lateral corner of the tibial plafond and attached to the fibula superior and medial to the inferior tip of the lateral malleolus. The deep fibres originated on the tibia, distal and medial to the centre of the superficial attachment, and inserted distally and medially to the superficial attachment on the fibula. The proximal aspect of the ITFL tibial attachment was located 45.0 ± 9.9 mm proximal to the plafond, while the distal aspect was found 11.1 ± 3.5 mm proximal to the tibial plafond (Fig. 5b). The cartilage facets of the syndesmotic tibiofibular contact zone were located along the lateral most aspect of the joint line at the intersection of the tibiofibular articulation, just lateral and slightly superior to the superior-lateral corner of the talar dome (Fig. 5b).

Discussion

The most important finding of this study was that the individual ligamentous and articular structures of the ankle syndesmosis were consistently identifiable with respect to anatomically defined and reproducible radiographic landmarks on both standard lateral and mortise radiographic projections. Additionally, measurements demonstrated excellent interobserver and intraobserver agreement for all structures of the syndesmosis on both lateral and mortise radiographic views. Quantitative attachment locations may be particularly useful in guiding surgical fixation in addition to facilitating continued development of anatomically-based surgical repairs and reconstructions.

The radiographic findings presented in this study correlated well with current anatomic descriptions in the literature. Bartonicek [1] reported that the superior extent of the ITFL was located 4–5 cm proximal to the joint line and the distal extent was located at 1–1.5 cm proximal to the tibial plafond. Subsequently, Ebraheim [10] reported corresponding measurements of 32.43 ± 4.11 and 8.10 ± 3.35 mm. Most recently, Williams et al. [44] reported the ITFL superior and inferior extents to be 49.4 [95 % confidence interval (CI) 45.4, 53.3] and 9.3 mm (95 % CI 8.3, 10.2) proximal to the central aspect of the tibial plafond. In the present radiographic investigation, the superior extent of the ITFL was located 45.9 ± 7.9 mm proximal and the distal extent was located at 12.4 ± 3.4 mm proximal to the tibial plafond on the lateral radiographic view. Similar distances were reported for the mortise view. Radiographic guidelines describing the location of the cartilage facets of the syndesmotic tibiofibular contact zone also correlated with anatomic descriptions. Williams et al. [44] reported the centre of the tibial cartilage facet to be 5.2 mm (95 % CI 4.6, 5.8) posterior to the anterolateral corner of the tibial plafond, while the present radiographic study reported the tibial cartilage facet to be 8.4 ± 2.1 mm posterior and superior to the anterior-most radiographically discernible point of the tibial plafond. The authors recognize that these landmarks and distances may not be directly comparable as anatomically visible and physically palpable landmarks may not directly coincide with what is radiographically identifiable; however, similarities between these measurements suggests that the anatomic structures were consistently identified across studies.

Likewise, agreement between previous anatomic descriptions and radiographic measurements presented here were also found for the commonly injured AITFL. Williams et al. [44] reported that the AITFL originated on the tibia 9.3 mm (95 % CI 8.6, 10.0) superior to the anterolateral corner of the tibial plafond and inserted on the fibula 5.8 mm (95 % CI 4.4, 7.3) proximal to the anteromedial (Wagstaffe’s) tubercle [44]. On the lateral radiographic view, the present study found the AITFL tibial attachment was 9.6 ± 1.5 mm superior and slightly posterior to the anterior corner of the tibial plafond and the centre of the fibular footprint was 4.4 ± 1.7 mm superior and posterior to the anterior-most point of the anterior fibular tubercle. These findings are evidence of strong agreement between anatomic and radiographic descriptions.

These radiographic guidelines have immediate and direct applications to anatomic reduction, surgical repair, or reconstruction following syndesmosis injuries. To date, surgical fixation and reconstruction techniques following syndesmotic injuries have been described for in vivo repairs as well as in cadaveric models [4, 9, 14, 21, 26, 28, 30, 32, 42, 45, 47]. In the case series reported, there are varying levels of success and a wide array of complications. In the case of acute syndesmosis injuries with instability, anatomic reduction via indirect transosseous fixation, either by syndesmotic screws or cortical button-suture constructs, is the current standard surgical practice [3, 24, 26, 32, 46].The current literature recommends that such fixation devices be placed between 2 and 5 cm proximal to the tibial plafond in line with the neutral tibiofibular orientation to avoid malreduction of the syndesmosis [32]. Despite these recommendations, malreduction is a frequently reported clinical complication, particularly with the use of syndesmotic screws [9, 25, 30, 42]. The incidence of malreduction with syndesmosis screw fixation has been reported to be as high as 52 % [12]. Fortunately, there is evidence that screw removal or screw breakage can lead to spontaneous reduction and improved symptoms in a high percentage of patients [15, 22, 40]. However, this suggests that ensuring initial anatomic reduction and fixation might lead to improved results including fewer broken screws or those requiring removal. The present study recommends that fixation screws or suture-button fixation devices be placed at a minimum of 12.4 mm and no more than 45.9 mm proximal to the tibial plafond on the lateral radiographic view to land within the footprint of the ITFL fibres and to ensure the safety of the synovial recess and articular surfaces. As recommended by previous studies, all devices should be inserted in line with the anatomic tibiofibular plane to avoid malreduction.

In addition to indirect fixation, various anatomic and non-anatomic reconstruction techniques have been described in the literature to address chronic instability, which also may be guided by the radiographic data presented in this study. Beumer et al. [4] initially described a technique in which an attenuated and elongated AITFL was retensioned through a proximal and medializing osteotomy of its tibial insertion. Grass et al. [14] subsequently described a modification of a peroneus longus ligamentoplasty in which a split peroneus longus tendon was threaded through a combination of three fibular and tibial canals to reconstruct the posterior, interosseous, and anterior ligaments of the syndesmosis. More recently, several authors have described free hamstring graft reconstructions including isolated AITFL reconstructions [45], combined AITFL/ITFL [28] and AITFL/PITFL [47], and complete syndesmosis triligamentous reconstructions [21]. Regardless of surgical technique, the radiographic guidelines defined in the present study could be utilized intraoperatively to guide the placement of reconstruction tunnels and fixation devices and to assess graft placement postoperatively. Specifically, the authors advocate that AITFL reconstruction tunnels be placed 9.6 mm superior and slightly posterior to the anterior-most radiographic aspect of the tibial plafond and 4.4 mm superior and posterior to the anterior fibular tubercle on the lateral view. Similar recommendations could be made for ITFL and PITFL reconstruction tunnels based on lateral and mortise measurements described in the present study. However, the authors would like to emphasize that such recommendations should be synthesized in conjunction with previously published gross anatomic data. Furthermore, suggested alterations in surgical technique in light of the presented radiographic data have not yet been evaluated biomechanically or clinically.

Clinical outcomes have often been reported to be satisfactory for both transosseous fixation and reconstruction; however, complications have also been reported including malreduction, residual diastasis, loss of range of motion (decreased dorsiflexion), and continued progression of degenerative joint disease [9, 14, 28–30]. Multivariate regression analysis of clinical outcomes has identified non-anatomic reduction (malreduction) as the only variable to independently influence patient outcomes [30]. Malreduction and failure to restore native joint contact mechanics is of particular clinical concern because previous biomechanical research has demonstrated that syndesmotic instability and widening of the ankle mortise, allowing for as little as 1 mm of relative lateral displacement of the talus, alters joint contact kinematics and reduces tibiotalar contact areas by as much as 42 % [36]. In addition to reduced contact areas, similar research has demonstrated that injuries resulting in altered tibiotalar contact mechanics significantly increase peak tibiotalar contact pressures [41]. It is believed that such non-physiologic contact areas and pressures can lead to subsequent chondral damage and arthritic changes [35]. The authors believe that the defined radiographic parameters presented here may facilitate fidelity to anatomic-based techniques and optimize the restoration of native syndesmosis joint kinematics postoperatively.

The authors acknowledge some limitations of the present study. This study utilized 12 cadaveric foot and ankle specimens. Given the relatively small sample size, the range of distances observed in this study may not represent the variability observed across a larger population. However, the number of specimens was comparable to previous radiographic landmark investigations [16, 20, 33, 34, 43]. Data were also comparable to previous anatomic literature [1, 10, 44]. In addition, specimens were generally obtained from older individuals that would fall outside of the typical age cohort that would undergo surgical syndesmotic fixation. However, specimens were screened for bone quality, osteophyte formation, joint space narrowing, and gross anatomic abnormalities. Based on these exclusion criteria, the authors are confident in the radiographic relationships established by this study. The authors also acknowledge that specimens were cut at the midshaft of the tibia and fibula, which may have altered the anatomic orientation of the syndesmosis; however, rigid screw fixation was utilized prior to removal of soft tissue to minimize any deviations from an anatomically accurate position. Finally, the reported measurements are two-dimensional quantitative descriptions of structures with three-dimensional relationships and therefore are subject to potential variability with rotation of the extremity. Therefore, careful adherence to the image acquisition protocol outlined in the materials and methods section is required to obtain results consistent with data presented in this study. Furthermore, the authors recommend that intraoperative navigation and surgical decision-making should always be made in conjunction with gross anatomic information detailing other anatomic soft tissue relationships.

This study provides a comprehensive description of the radiographic anatomy of the ankle syndesmosis, including ligament attachments and articular surfaces. This information will assist in the interpretation of radiographic assessments of the syndesmosis from diagnosis through post-operative follow up. Such guidelines may be particularly useful in more difficult revisions or cases with significant concomitant injury where other means of assessment and navigation may not be easily applied.

Conclusions

In the present descriptive laboratory study, qualitative and quantitative radiographic parameters characterizing relevant ligament attachment sites and cartilage surfaces of the ankle syndesmosis were defined with excellent reliability and reproducibility. In conjunction with current anatomic data, these radiographic guidelines will augment current clinical radiographic diagnostic techniques, improve preoperative planning, assist with intraoperative identification of native anatomy, and facilitate objective postoperative assessment of anatomic-based reduction, repair, and reconstruction techniques.