Chapter No. 1

1. INTRODUCTION

Injuries to the distal tibiofibular syndesmosis are complex and remained controversial with regard to diagnosis and management. In United Kingdom, ankle fractures are the most common fracture among patients aged between 20 and 65 with the annual incidence reported as 90,000 (1). Twenty percent20% of ankle fractures requireing internal fixation (2), and or 10% of all ankle fractures are associated with syndesmosis disruption (3). Syndesmotic injuries have also been reported in the absence of fracture and sometime called as “high ankle sprain”with incidence reported somewhere between 1% and 11% of all ankle fractures or 0.5% of all ankle sprains (4-6). Despite the considerable tremendous amount of work load these injuries provide for orthopaedic surgeons, there is no consensus regarding the optimal treatment of these injuries, resulting and sometime results in under or over treatment of syndesmotic injuries, especially those without fibular fracture. It is therefore important to understand the anatomy, biomechanics and the mechanism of injuries involving the tibiofibular syndesmosis.

1.1. Anatomy

The inferior tibiofibular joint is a syndesmotic joint formed by two bones and four ligaments. The distal tibia and fibula form the osseous part of the syndesmosis held together by four ligaments providing stability that is integral for proper functioning of the ankle joint (6-8). These ligaments include the anterior inferior tibiofibular ligament (AITFL), the posterior inferior tibiofibular ligament (PITFL), the transverse tibiofibular ligament and the interosseous ligament.

At the apex of syndesmosis, the interosseous border of tibia bifurcates caudally into an anterior and posterior margin. The anterior margin ends in the antero-lateral aspect of the tibial plafond called the anterior tubercle (Chaput's tubercle). The posterior margin ends in the posterolateral aspect of the tibial plafond called the posterior tubercle. The anterior and posterior margins of the distal tibia enclose a concave triangular notch called insisura fibularis, with its apex 6-8 cm above the level of the talocrural joint (9-11). The anterior tubercle is more prominent than the posterior tubercle and protrudes further laterally and overlaps the medial two thirds of the fibula (9-11). The fibular part of the syndesmosis is convex and matches with its tibial counterpart. The crista interossea fibularis, i.e. the ridge on the medial aspect of the fibula, also bifurcates into an anterior and posterior margin and forms a convex triangle that is located above the articular facet on the lateral malleolus. The base of the fibular triangle is formed by the anterior tubercle (Wagstaffe-Le Fort tubercle) and the, almost negligible, posterior tubercle (9). Shape of insisura fibularis varies among individual. Elgafy et al (12) described two main morphological patterns in their study of 100 normal ankle syndesmoses. In 67% the insisura was deep, giving the syndesmosis a crescent shape while in 33% it was shallow, giving the syndesmosis a rectangular shape (12).

The anterior inferior tibiofibular ligament AITFL runs obliquely from anterior tubercle of distal tibia to anterior tubercle of fibula [Fig. 1.1]. AITFL consists of multifascicular bundle of fibers that run obliquely downwards and laterally and prevents excessive fibular movement and external talar rotation (13). The AITFL is the first ligament to fail in external rotation injuries (9). Posterior inferior tibiofibular ligament PITFL is a strong ligament. It originates from posterior tubercle of distal tibia and runs obliquely downwards and laterally to the posterior lateral malleolus (14) [Fig. 1.2]. PITFL works along with AITFL to hold the fibula tight in insisura fibularis of the tibia. The lower part of the PITFL runs more horizontally and is considered as a separate anatomical entity called transverse ligament. The transverse ligament is a thick, strong structure with twisting fibers. It passes from the posterior tibial margin to the posterior margin of malleolar fossa of distal fibula. The location of the transverse ligament below the posterior tibial margin creates a posterior labrum, which deepens the articular surface of the distal tibia and helps to prevent posterior talar translation [Fig. 1.2]. The interosseous tibiofibular ligament is a thickening of lower most part of interosseous membrane and consists of numerous short, strong, fibrous bands which pass between the contiguous rough triangular surfaces of the distal tibia and fibula and form the strongest connection between these bones, providing stability to talocrural joint during loading. The ligament is thought to act like a spring, allowing for slight separation between the medial and lateral malleolus during dorsiflexion at the ankle joint and thus for some wedging of the talus in the mortise (9).

Ogilvie-Harris et al (15) studied the relative importance of each of the ligaments in the distal tibiofibular syndesmosis using 8 fresh-frozen cadaver specimens to evaluate the percentage of contribution of each ligament during 2 mm of lateral fibular displacement. The anterior inferior tibiofibular ligament provided 35%; the transverse ligament, 33%; the interosseous ligament, 22%; and the posterior inferior ligament, 9%. Thus, more than 90% of total resistance to lateral fibular displacement is provided by 3 major ligaments. Injury to one or more of them result in weakening, abnormal joint motion, and instability.

1.2. Biomechanics

The primary movements at the ankle joint include dorsiflexion and planterflexion. The normal ankle allows approximately 15o to 20o of active dorsiflexion which may be increased to 40o passively and between 45o to 55o of plantar flexion (16). The superior surface of the talus is wedge shaped and wider anteriorly than posteriorly with an average difference of 4.2 mm (17). During dorsiflexion, the wider anterior portion of the talus ‘‘wedges'' between the medial and lateral malleoli, and much of the mortise becomes occupied (6). Up to 6o of talar external rotation occurs during ankle dorsiflexion and the talusit rotates internally and supinates slightly during plantar flexion, as a result of its conical and wedged shape (17-19). During normal ankle motion, some movement occurs normally at the distal tibiofibular syndesmosis. Although ankle syndesmosis is a tightly held fibrous joint it allows 1 to 2 mm of widening at the mortise as the foot is moved from full plantar flexion to full dorsiflexion. This widening of mortise occurs partly as a result of 3o to 5o of fibular rotation along its vertical axis during plantar flexion and dorsiflexion (6, 18, 20).

When fixing ankle fractures, it is vital necessary to restore normal anatomic relations of distal tibiofibular syndesmosis, as slight discrepancy can lead to significant change in biomechanics and sub optimal long term results. Ramsey and Hamilton (21) demonstrated that as little as 1 mm of lateral shift of the talus in the ankle mortise resulted in a 40% loss of tibiotalar contact surface area and increase in contact stresses. Similar findings were also confirmed by another recent study by Lloyd et al (22) in 2006. Taser et al (23) showed using three-dimensional computed tomographic (CT) reconstructions that a 1 mm separation of the syndesmosis can lead to a 43% increase in joint space volume.

1.3. Mechanism of Injury

The 3 proposed mechanisms of ankle syndesmotic injury include external rotation of the foot, eversion of the talus and hyper dorsiflexion (6, 24). External rotation injuries result in widening of the mortise as the talus is forcefully driven into external rotation within the mortise. Forceful eversion of the talus also results in widening of the mortise. These mechanisms are most common in sports like football and skiing. Hyperdorsiflexion injuries are seen in jumping sports and also result in widening of mortise when wider anterior part of the talus dome is forcefully driven into the joint space. In all cases, the fibula is pushed laterally and if the forces are strong enough, leads to diastasis of ankle syndesmosis (24-30).

Lauge-Hansen (31) classified the ankle fractures according to the mechanism of injuries. This classification system was based on cadaveric study and takes into account the position of foot at the time of injury and the deforming force. According to this syndesmotic disruption most commonly occurs in “Pronation-External Rotation” (PER) injuries. Depending on the severity of the force applied, this abnormal movement will result in rupture the deltoid ligament or fracture the medial malleolus in its first stage, with subsequent injury to the syndesmotic ligaments and the interosseous membrane, and finally a spiral fracture of the fibula above the level of syndesmosis (31, 32). Most of the complete syndesmotic disruptions are associated with Weber C fracture with smaller proportion having Weber B fracture with widening of the mortise and, occasionally, a Maissonneuve fracture (33). Syndesmotic diastesis rarely occurs in isolation without bone injury and poses a diagnostic challenge. These injuries are sometime referred as “high syndesmotic sprain” (4, 27, 34).

1.4. Diagnosis

Diagnosis of syndesmotic injury can sometime be challenging and depends on high index of suspicion, taking into consideration, the mechanism of injury and the clinical findings and confirming with radiological assessment or examination under anaesthesia. Several clinical tests have been described in literature but lack high predictive value in acute cases as it might be difficult to perform these tests because of excessive pain in acute situations. Some examples of these tests include Squeeze test (34), Point test (35), External rotation test (32, 35) and Fibular translation test (32, 36).

Radiographs are important in diagnosis of tibiofibular syndesmotic diastasis. Three radiographic parameters have been described based on anterior-posterior and mortise views but controversy exist among researchers with regard to the optimal parameter for accurate diagnosis. The “tibiofibular clear space” is defined as the distance between the lateral border of the posterior tubercle and the medial border of the fibula. The “tibiofibular overlap” is the distance between the medial border of the fibula and the lateral border of the anterior distal tibial tubercle and the “medial clear space” is the distance between the articular surface of medial malleolus and the adjacent surface of talus (32, 37). Harper et al (38) radiographically evaluated normal tibiofibular relationship in 12 cadaver lower limbs and based on a 95% confidence interval, demonstrated following criteria as consistent with a normal tibiofibular relationship: (1) a tibiofibular "clear space" on the anterior-posterior and mortise views of less than approximately 6 mm; (2) tibiofibular overlap on the anterior-posterior view of greater than approximately 6 mm or 42% of fibular width; (3) tibiofibular overlap on the mortise view of greater than approximately 1 mm. The study concluded that the width of the tibiofibular "clear space" on both anterior-posterior and mortise views appeared to be the most reliable parameter for detecting early syndesmotic widening and medial clear space greater than a superior clear space is indicative of deltoid ligament injury (38). The accuracy of these measurements has been questioned in several studies. Beumer et al (39) demonstrated that these measurements are greatly influenced by the positioning of ankle while taking radiographs. Similar findings were confirmed by Nelson et al (40) and Pneumaticos et al (41) except that the later study reported that the tibiofibular clear space did not change significantly by rotation of ankle (41). CT and MRI scanning are more sensitive than radiography for detecting minor degrees of syndesmotic injury and provide an important diagnostic tool in suspicious cases (7, 42).

1.5. Treatment of Syndesmosis diastasis and review of literature

Injuries to distal tibio-fibular syndesmosis are complex and require accurate reduction and fixation for optimal outcome (43, 44) but the choice of fixation still remained controversial. Kenneth et al (45) studied the effect of syndesmotic stabilization on the outcome of ankle fractures in 347 patients at a minimum follow up of 1 year and concluded that patients requiring syndesmotic stabilization in addition to the malleolar fixation had poorer outcome as compared to patients requiring only malleolar fixation.

Although, the use of metal screw has been the most popular means of stabilizing the syndesmosis (32), controversy exists with regard to the size and number of screw, number of cortices engaged, level of screw placement above the tibial plafond, need for routine removal and the timing of the screw removal (46-48). Beumer et al (49) in their cadaveric study, reported no difference in fixation of the syndesmosis when stainless steel screws were compared to titanium screws through three or four cortices. Hoiness et al (46) conducted a randomised prospective trial comparing single 4.5 mm quadricortical screw with two 3.5mm tricortical screws for ankle syndesmosis injuries in 64 patients. The study showed improvement in early function in the tricortical group, but after one year there was no significant difference between the groups in their functional score, pain or dorsiflexion (46). Further report on the same study group with 8.4 years average follow up did not show any significant difference in clinical outcome (50). Moore et al (51) also reported similar functional outcome with either three or four cortical fixation using 3.5 mm screws with slightly higher trend toward loss of reduction in tricortical group. Although there is no clinical consensus regarding number and size of the screws, biomechanical studies have shown that two screws are mechanically superior to single screw (52). There is no significant difference between 3.5 mm and 4.5 mm syndesmosis screw when used as tricortical screw (48) but when used as quadricortical screw 4.5 mm screw showed higher resistance to shear stress than 3.5 mm screw (53). Routine removal of syndesmosis screw is another controversial issue. Some authors advocate routine removal before starting full weight bearing as screw provides rigid fixation of syndesmosis where micromotion occurs normally and can therefore lead to screw loosening or fatigue failure (54-57). Miller et al (58) demonstrated improved clinical outcomes following syndesmosis screw removal in a series of 25 patients. Manjoo et al (59) retrospectively reviewed 106 patients treated with syndesmosis screw. Seventy-six returned for follow up. The study concluded that intact screw was associated with a worse functional outcome as compared with loose, broken or removed screws. However there were no differences in functional outcomes comparing lose or broken screws with removed screws (59). Both these studies had inherent limitations including of retrospective studies study design and lack of a the control group.

Malreduction of tibiofibular syndesmosis has been reported as a significant problem with screw fixation and is an independent predictor of functional outcome (44). Gardner et al (60) reported 52% of malreduction of syndesmosis in weber C fractures treated with screw fixation.

Bioabsorbable screws haves also been used as an alternative to metal screws to avoid hardware related complications and haves demonstrated equal effectiveness in fixation of diastesis (61-63). However, these implants did not gain popularity because of concerns including osteolysis, foreign-body reaction, late inflammatory reaction and osteoarthritis due to polymer debris entering the joint (64-67).

The Arthrex Tightrope is a relatively new surgical implant based on the suture endobutton design. It is a low profile system comprised of a No. 5 FiberWire® loop which, tensioned and secured between metallic buttons placed against the outer cortices of the tibia and fibula, provides physiologic stabilization of the ankle mortise and obviates the need for a second procedure for removal, therefore late diastasis is unlikely (68). Biomechanical testing and clinical trials have shown equivalent strength and improved patient outcome with the tightrope technique (69, 70). In 2005 Thornes et al (71) performed a clinical and radiological comparison of 16 patients treated with suture-button techniques with similarand a similar cohort of patients treated with syndesmosis screw fixation. Patients in suture button group demonstrated significantly better American Orthopaedic Foot and Ankle Society (AOFAS) score and returned to work earlier than screw group. As with any novel technique, the follow-up reported in the literature is short and the number of cases are limited [Table 1]. The largest case series so far, has reported the outcome in 25 cases patients (72, 73). Although initial series did not report any complications, some cases of implant removal have been reported in more recent literature because of soft tissue irritation. In a series of 16 patients, two tightropes were removed, one due to infection, and the other due to soft-tissue irritation (74). Willmott et al (75) reported 2 cases of tightrope removal because of soft tissue inflammation, out of 6 patients treated with ankle tightrope (33%). One of them was removed because of inflammation over medial button. Coetzee et al (76) in their results of a prospective randomized clinical trial also reported removal of one tightrope because of infection, out of 12 cases. In a most recent series of 24 cases DeGroot et al (77) reported removal of hardware in 6 patients due to soft tissue complication. They also reported subsidence of endo-button due to osteolysis in adjacent bone in 4 cases but did not have any effect on clinical outcome as it was a late occurrence. There were also 3 cases of heterotopic bone formation in this series.

Table 1 : Studies reporting o n clinical outcomes and complications of Tightrope fixation .

Authors

Year

Number

Followup

(months)

Time to FWB

(Weeks)

AOFAS

score

No. of

complication

Seitz et al (69)

1991

12

38

-

-

0

Thornes et al (71)

2005

16

12

-

93

0

Mcmurray et al (74)

2008

16

5

6

87

2

Cottom et al (72)

2008

25

10.8

5.5

50.6*

0

Willmott et al (75)

2009

6

5.3

6

-

2

Coetzee et al (76)

2009

12

27

-

94

1

DeGroot et al (77)

2011

24

20

5.7

94

6

AOFAS; American Orth opaedic Foot and Ankle Society Score.

*  Cottom et al used a modified AOFAS score with maximum score of 63.

Despite satisfactory short term clinical outcomes, few complications have also been reported related to soft tissue irritation and also there is a concern that tightrope might be inferior to screw in maintaining the syndesmosis. So far, the literature is limited with regard to tightrope fixation and the issue of malreduction has not been properly investigated. Radiological measurements in most of the studies are performed on radiographs. It has been previously noted that radiographic measurements are influenced by the rotation of ankle and therefore not accurate. Thornes et al performed axial CT scan on 11 of 16 patients treated with tightrope at 3 months and did not find any malreduction (71). CT scans were performed only after 3 month of surgery and none of the patient in control group had a CT scan and therefore undermines the significance of this part of their study. Significant malreduction of tibiofibular syndesmosis has been reported in literature for patients treated with syndesmosis screw (50, 60). As malreduction of syndesmosis is the most important independent predictor of long term functional outcome we aim to fill the gap in literature regarding tightrope's ability to maintain syndesmosis integrity in longer term.

Table 1.1 : Studies reporting o n clinical outcomes and complications of Tightrope fixation .

Authors

Year

Number

Followup

(months)

Time to FWB

(Weeks)

AOFAS

score

No. of

complication

Seitz et al (69)

1991

12

38

-

-

0

Thornes et al (71)

2005

16

12

-

93

0

Mcmurray et al (74)

2008

16

5

6

87

2

Cottom et al (72)

2008

25

10.8

5.5

50.6*

0

Willmott et al (75)

2009

6

5.3

6

-

2

Coetzee et al (76)

2009

12

27

-

94

1

DeGroot et al (77)

2011

24

20

5.7

94

6

AOFAS; American Orth opaedic Foot and Ankle Society Score.

*  Cottom et al used a modified AOFAS score with maximum score of 63.

1.6. Aims and Objective

The primary A aim of this study is to compare the accuracy and maintenance of syndesmotic reduction using tightrope technique and syndesmosis screw fixation and their consequences on clinical outcome.

Population (P) - Adult patients with acute fixation of ankle syndesmosis.

Intervention (I ) - Tightrope fixation of ankle syndesmosis.

Comparison (C) - Syndesmosis screw fixation.

Outcome (O) - Accuracy of syndesmotic reduction, based on axial CT scan.

Chapter No. 2

2. PATIENTS AND METHODS

We conducted a cohort study to assess the radiological and clinical outcomes of patients after treatment of ankle injuries involving distal tibiofibular syndesmosis. Two different methods of syndesmosis fixation were compared (standard transosseous syndesmosis screw fixation and a relatively new, Tightrope fixation technique) for the accuracy and maintenance of syndesmosis reduction and its correlation with the functional outcome scores after at least 18 months following the index procedure. The accuracy of syndesmosis reduction was measured primarily on axial Computed Tomographic (CT) scans and anterio-posterior (AP) radiographs of ankles using uninjured contralateral ankle as a control.

The study was conducted in department of Trauma and Orthopaedics and the department of Radiology in Our Lady of Lourdes Hospital, Drogheda, Republic of Ireland after approval by the Institutional Review Board (appendix i). The patients were recruited using trauma theatre database. The data regarding all patients treated for ankle injuries was reviewed.

The inclusion criteria were as follows:

adults (> 18 years) with acute ankle syndesmosis injury

willing to give informed consent to participate in the study

, fixation of the injuryed over a 2 years period from July 2007 to June 2009 provided they did not fit into the exclusion criteria.

The exclusion criteria set out for this study included:

P patients with open fracture,

I i ndividuals with diabet es ic or neuropathic arthropathy,

M multi trauma patients and

P patients who had a previous injury or surgery on the contra-lateral ankle as those could not be used as a control.

Pregnancy was included in exclusion criteria B because of radiation exposure in this study. “pregnancy” was also mentioned as exclusion criteria.

i I ndividuals unwilling to consent to the study

Patients were treated by six Orthopaedic consultants in a single trauma unit using two different techniques for syndesmosis fixation including traditional screw and tightrope fixation technique. Three consultants used screw fixation while the other three consultants used tightrope technique for all of their patients requiring syndesmosis fixation irrespective of age, sex and the type of associated fractures. The diagnosis of tibiofibular diastasis was based on careful clinical examination, consideration of the fracture pattern and radiographic parameters including widening of medial clear space (MCS), increased tibiofibular clear space (TFCS) and reduced tibio-fibular overlap (TFOL) preoperatively; and intraoperative confirmation under fluoroscopy using “external rotation stress test” and “hook test” in which fibula was pulled laterally after fixation of fracture using a bone hook and widening of syndesmosis was observed using image intensifier. Concomitant fractures of fibula and medial malleolus were fixed according to standard AO principles. Ankle syndesmoses were stabilized with either “Transosseous Screw” or “Tightrope” depending on the consultant's preference. All patients were immobilized in below knee plaster back slab for two weeks followed by non-weight bearing cast for another four weeks. Casts were removed in after six weeks time and patients were referred for physiotherapy and allowed full-weight bearing as tolerated. Patients were followed up in clinic at 2 weeks, 6 weeks and then after 3 months. Patients were finally reviewed in January 2011 for the collection of study data. Patients who consented for the research participationto this study underwent a clinical examination by an independent clinician who was blinded for the type of syndesmosis fixation. Two functional scoring systems were used to assess clinical outcome, including a clinician reported American Orthopaedic Foot and Ankle Society (AOFAS) scoring system (78) and a patient reported Foot and Ankle Disability Index (FADI) score (79). Radiographic assessment included anterior-posterior radiograph of both the ankles together and an axial CT scan of both the ankles together at 1 cm above the tibial plafond. All the CT scans were performed by single, senior CT Radiographer using same specifications.  All patients were scanned supine in the axial plane with no gantry tilt.  Survey CT scan image was obtained first instead of scanning the whole ankle, to reduce the radiation dose. The area of ankle syndesmosis was scanned using single slice CT scan. The thickness of the CT slice was 3.8 mm and was centred at 12 mm from the tibial plafond as measured on the survey scan image. This sSingle slice scan provided two axial images, one at approximately 1 cm from the tibial plafond and other at 1.4 cm approx [Fig. 2.1]. This technique was adopted in order to reduce the radiation exposure to the patient without compromising the quality of the scans and the axial images thus obtained correspond to the same level as used for the measurements on radiographs i.e. 1 cm above tibial plafond.

2.1. Outcome Variables

The “accuracy of syndesmosis reduction” on axial CT scan was considered as primary outcome variable to compare the two different treatment options. The criterion for malreduction of syndesmosis was set at > 2 mm of difference in the width of syndesmosis as compared with the normal contralateral ankle when measured on the axial CT scan. The width of posterior part of syndesmosis joint space was measured for the purpose of this comparison as this measurement correspond to the tibiofibular clear space on AP radiographs. The criterion was set at 2 mm in accordance with previous literature (60) and the assumption that this difference will result in sufficient level of joint incongruity which may lead to increased contact pressures in ankle joint and the risk of early degenerative changes (21, 22). Elgafy et al (12) reported that the average width of syndesmosis posteriorly is 4 mm with standard deviation of 1.19 mm. As this area corresponds to the tibiofibular clear space on AP radiographs and > 6 mm of tibiofibular clear space is considered abnormal, the criterion of > 2 mm would be justified. 

Syndesmosis integrity was also assessed on AP radiographs of ankle, using parameters including “tibiofibular clear space (TFCS < 6 mm)”, “tibiofibular overlap (TFOL >6 mm)” and “medial clear space (MCS < 5 mm)”.

Clinical outcomes were assessed using two functional scores, time to full weight bearing and rate of complications. Functional scoring systems include American Orthopaedics Foot and Ankle Society (AOFAS) score (appendix ii) which has been widely used in previous ankle studies. It is a clinician reported scoring system which looks at the pain, functional status, alignment and range of motion of foot and ankle. Foot and Ankle Disability Index (FADI) score (appendix iii) is a patient reported functional scoring system and looks at pain and various functional activities. Both the scores range from 0 to 100 with higher scores indicating better function.

In the statistical analysis, factors considered potential confounders were patient's age and the durationtime since surgery. These confounders were adjusted using regression analyses.

2.2. Data Collection and Measurements

Demographic data of the patients and the data regarding the mechanism of injury, type of fractures and the type of fixation were extracted from patient's clinical notes.

Radiographic parameters of syndesmosis integrity were measured on preoperative and the latest AP ankle radiographs 1 cm proximal to the tibial plafond. The “tibiofibular clear space” is defined as the distance between the lateral border of the posterior tibial tubercle and the medial border of the fibula. The “tibiofibular overlap” is the distance between the medial border of the fibula and the lateral border of the anterior distal tibial tubercle and the “medial clear space” is the distance between the articular surface of medial malleolus and the adjacent surface of talus (32, 37).

The width of syndesmosis was measured on axial CT scan for both operated and normal ankles simultaneously. Measurements were performed on axial scan 1 cm proximal to the tibial plafond as described earlier to provide measurements that are comparable to those obtained on standard radiographs. The distal fibula shows considerable variation with regard to the prominence of the borders. Four borders of fibula have been mentioned in anatomy textbooks including anterior, posterior, medial and interosseous border [Fig 2.2]. As the cross-sectional anatomy of distal tibia is more constant than fibula we used anterior and posterior tibial tubercle as our reference points for the measurements of anterior and posterior syndesmosis width. Although anterior width of syndesmosis was not used for comparison of malreduction, it was measured to evaluate normal anatomic variations in uninjured ankle syndesmosis. Two measurements were performed. Anterior width was measured from anterior tibial tubercle to the nearest point on the fibula. Similarly, the posterior width was measured from posterior tibial tubercle to the nearest point on the medial boarder of fibula [Fig. 2.3]. Measurements were performed by an independent musculoskeletal rRadiologist on using digital software on CT work station. CT measurements were performed twice at an interval of 2 weeks without the knowledge of previous readings, to assess the intra-observer agreement.

Clinical assessment was performed by an independent clinician who was blinded for the type of syndesmosis fixation. AOFAS and FADI scores were completed at this review.

As this was a non-randomised study there were possibilities of bias and every effort was made to reduce the bias. All consecutive patients who fulfilled the eligibility criteria were invited for participation in the study to reduce the selection bias. It was confirmed retrospectively that different methods of syndesmosis fixation were assigned to patients only on the basis of surgeons preferred choice, irrespective of age, sex or type of associated fracture. This means thatIn essence, patients admitted on certain days of the week were fixed with syndesmosis screw and patients attending on the remaining days were treated with tTightrope fixation technique. Measurements on the CT scan and radiographs were performed by an independent radiologist. As blinding was not possible, measurements were performed twice at an interval of 2 weeks to assess intra-observer reliability of measurements. Finally, the clinical assessment was performed by an independent clinician not directly involved in the study and was blinded to the type of fixation. This was important to reduce assessor's or interviewer's bias.

2.3. Sample Size

Sample size was calculated on stata 11.1 for comparison of two means, using measurements of normal syndesmosis on CT scan as reported by Elgafy et al (12) . Using mean of 4 mm and standard deviation of 1.19 and considering 2mm as clinically significant difference gives minimum of 10 cases in each group for 90% power. Although 2 mm difference is used for detection of malreduction in individual patient, there might not be a mean difference of 2 mm. Therefore we calculated the sample size for one standard deviation difference from the normal mean value which requires 22 cases in each group.

A sample size calculation was performed based on the primary outcome measure i.e. measurements of normal syndesmosis on CT scan as reported by Elgafy et al (12) Elgafy et al . The formula used to determine the number of participants required in the study involved the prediction of the standard deviation ( σ ) for normal CT measurements and an anticipated significant clinical change or deviation from normal CT measurements of the ankle (Δ) [Fig. 2.4] . The value for the σ was obtained from the paper by Elgafy et al (12) Elgafy et al . Although 2 mm difference is widely used for detection of malreduction in individual patient s , the difference is often not 2 mm in individuals presenting with problems. Therefore we considered 1 millimetre as a clinically significant difference (Δ) so that the final power of the study is not undermined . The value of the constant ‘K', 7.8, was dictated by the significance level chosen for the study, in this case a two-sided significance level of 5% with an 80% chance of detecting a treatment effe ct. Based on a two group comparison, power calculations indicated that a minimum of 46 participants were required to detect a change of 1mm on the CT measurements at a two-sided significance level of 5% and a power of 80%, assuming a σ of 1.19 points. This analysis was confirmed using Stata 11.1 ® statistical software.

Number of participants required in each of the comparison groups must be greater than the value calculated using the following formula

2 (Constant K) ( σ of the normal CT measurements) 2

(What is considered to be a clinically significant change in CT measures) 2

2 (7.8 for two sided test with significance level of 0.05) (1.19) 2

(1) 2

2 (7.8) (1.42)

1

23 participants per group

Therefore in order to detect a clinically significant change of 1 mm deviation on the CT measurement, a minimum of 46 participants were required in total

2.4. Statistical Analysis

Statistical analysis was performed on Stata 11.1®. Demographics were compared for the two groups using mean values and proportions. Mean, standard deviation, ranges and confidence interval (CI) were calculated for the continuous variables including age, follow-up, time to full weight bearing, radiographic and CT parameters and functional outcome scores. Mean values were calculated for the radiographic and CT parameters for both operated and normal ankle in two groups separately and compared using t-test within each group for measurement of statistical significance. Difference in the width of syndesmosis between normal and operated side were calculated and compared using unpaired t-test with p-value < 0.05 to be considered statistically significant. 2x2 table was formulated for categorical variables including malreduction of syndesmosis and complications and were analysed to calculate relative risk and statistical significance using fisher's exact test. Potential confounders including age and duration since surgery were accounted for using regression analysis when analysing the correlation of syndesmosis malreduction with functional outcome scores. 

As our primary outcome variable was accuracy of syndesmosis reduction based on syndesmotic width measured on axial CT scans, we also assessed intra-observer agreement for CT measurements using “intra-class correlation coefficient” (ICC) (80). The values for ICC range from 0.0 to 1.0 and can be interpreted as follows: 0-0.20 indicates poor agreement: 0.21-0.40 indicates fair agreement; 0.41-0.60 indicates moderate agreement; 0.61-0.80 indicates strong agreement; and >0.80 indicates almost perfect agreement.

2.5. Data protection

The study was approved by the institutional review board and due consideration was given to data protection rules (appendix ivii). Patients were fully informed about the purpose of the study and outcome measures using an information leaflet (appendix iv). All patients included in the study voluntarily signed an informed consent (appendix vi). Patient's data was stored on electronic database using unique identification code making it completely anonymous for analysis and storage purpose. Research data will be retained for a minimum of 5 years after the publication of the research. 

Chapter No. 3

R ESULTS

3. 1. Participants

Of the 228 consecutive patients operated for ankle fractures during the study period of 2 years, 167 patients did not have any syndesmosis injury.  Sixty one patients had associated syndesmosis injuries and were potentially eligible for study inclusion. Six of those 61 patients were excluded on the basis of study exclusion criteria. One patient died before recruitment, 2 were visitors from abroad and were  uncontactable, 2 had compound injury and 1 had bilateral injuries to his ankles. Fifty five patients were finally eligible for the study and invited for participation. Forty nine consented for the study, 5 refused to participate in the study as they did not have any problem and thought a review is unnecessary. One more patient moved abroad by that time. Out of 49 patients who consented for the study, 3 more were not able make it to the appointment because of work commitments leaving 46 patients for final analysis who attended for final follow up and CT scan.

3.2. Patient's demographics and injury classification

Forty-six patients finally attended for the review, 23 in tightrope group and 23 in screw group. Mean age was 41.65 years (range 24 - 69 years) and 39.82 years (range 18 - 65

Table 3.1: Comparison of patient's demographics and injury pattern between two groups

Tightrope Group

Syndesmosis Screw Group

Total number

23

23

Gender

Male

17 (74%)

16 (70%)

Female

06 (26%)

07 (30%)

Age(years)

41.65 (24 - 69)

39.82 (18 - 65)

Side

Right

08 (35%)

10 (43%)

Left

15 (65%)

13 (57%)

Mechanism of injury

Sports

5 (21.7%)

6 (26.1%)

Fall from height

6 (26.1%)

3 (13.1%)

Trip and fall

9 (39.1%)

7 (30.4%)

Slipped on ice

3 (13.1%)

5 (21.7%)

Dancing

0

2 (08.7%)

Classification

Weber B (SER)

02 (08.7%)

02 (08.7%)

Weber C (PER)

13 (56.5%)

15 (65.2%)

Maisennouve

08 (34.8%)

06 (26.1%)

Number of fixations

Single

16

20

Double

7

3

years) respectively in each groups. There were 17 (74%) male and 6 (26%) female in tightrope group while screw group had 16 (70%) male and 7 (30%) female. Right ankle was operated in 8 (35%) in tightrope group and 10 (43%) in screw group while left ankle was operated in 15 (65%) and 13 (57%) patients respectively.  There were 2 weber B fractures, 13 weber C and 8 Maisennouve fractures in tightrope group while 2 weber B, 15 weber C and 6 Maisennouve fractures in screw group. In tightrope group single tightrope was used in 16 patients while two tightropes were used in 7 patients. In screw group 20 patients had single syndesmosis screw while 3 patients required double screw fixation.

Mean follow up was 30.2 months (range 18 - 41 months) in tightrope group and 29 months (range 18 - 41 months) in syndesmosis screw group.

3.3. Computed Tomographic measurements

Measurements for the normal tibiofibular syndesmosis are summarized in [Table 3.2]. Mean tibiofibular width in normal ankles were 2.85 mm (range 1.9 - 4.4mm), anteriorly and 4.03 mm (2.2 - 6.3mm), posteriorly. In men the mean anterior width was 2.7 mm and posterior width was 4.12 while in women mean width was 3.23 mm anteriorly and 3.81 mm posteriorly. The measurements were performed twice in random order at least two weeks apart and analysed for intra-observer agreement. The intra-class correlation coefficient value was 0.91 for the two measurements.

Comparison of syndesmosis width between normal and operated ankle showed mean values of 4.04 + 0.95 mm for normal side and 4.37 + 1.12 mm for operated side in tightrope group ( p = 0.30, t-test). In syndesmosis screw group the mean width of syndesmosis was measured as 4.02 +0.87 mm on the normal side and 5.16 + 1.92 mm on the operated side ( p = 0.01, t-test) [Table 3.3] [Fig. 3.2].

Table 3.2: Mean values of the width of normal syndesmosis

Mean, standard deviation and range

Anterior width (mm)

Posterior width (mm)

Total

Mean

2.85

4.03

n = 46

Standard deviation

0.75

0.9

Range

1.9 - 4.4

2.2 - 6.3

Male

Mean

2.7

4.12

n = 33

Standard deviation

0.68

0.91

Range

1.9 - 4.4

2.2 - 6.3

Female

Mean

3.23

3.81

n = 13

Standard deviation

0.8

0.87

Range

2.1 - 4.4

2.7 - 5.6

Table 3.3: Comparison of syndesmosis width between normal and operated ankle

in two groups

Normal ankle

Operated ankle

p value (t-test)

Tightrope group

4.04 + 0.95 mm

4.37 + 1.12 mm

P = 0.30

n = 23

(2.2 - 6.0) 

(2.5 - 6.4)

Screw group

4.02 + 0.87 mm

5.16 + 1.92 mm

p = 0.01

n = 23

(2.7 - 5.6) 

(2.1 - 10.3) 

All values are mean values in mm + standard deviation (SD) and (ranges). P value <

0.05 is considered statistically significant. 

Table 3.4: Malreduction of syndesmosis between tightrope and screw group

Malreduction

No

Yes

Total

Tightrope group (n = 23)

23

0

23

Screw group (n = 23)

18

5 (21.73%)

23

Total

41

5

46

Malreduction was diagnosed on the bases of pre defined criteria of > 2 mm difference

from the normal side. p <0.05 Fisher's exact test.

The primary outcome measure, malreduction of syndesmosis was diagnosed on the basis of pre-defined criteria of > 2 mm difference from the normal side [Table 3.4]. There was no case of malreduction in tightrope group as compared to 5 (21.7%) cases of malreduction out of 23 cases of syndesmosis screw fixation (p <0.05, Fisher's exact test) [Fig. 3.3]. Risk of malreduction was 21.7% higher in screw group than tightrope group.

3.4. Radiographic measurements

Radiographic parameters of syndesmosis integrity were measured on standard AP radiograph of ankle 1 cm above the tibial plafond. Mean pre operative and post operative values are compared in [Table 3.5] [Fig. 3.4]. Mean post operative medial clear space (MCS) was 3.36 + 0.5 mm in tightrope group and 3.23 + 0.6 mm in syndesmosis screw group (p = 0.48). In tightrope group the mean post operative tibiofibular clear space (TFCS) was 4.04 + 0.8 mm as compared to 5.0 + 1.8 mm in screw group (p < 0.05) while mean tibiofibular overlap (TFOL) was 8.21 + 2.0 mm and 7.47 + 2.0 mm respectively (p = 0.22).

Table 3.5: Pre and post operative radiographic parameters of syndesmosis integrity

Tightrope group

Syndesmosis screw group

p value

Medial clear space

Pre-Op

5.86 + 2.3 mm (3 - 15)

6.67 + 1.7 mm (4 - 10)

Post-Op

3.36 + 0.5 mm (2 - 4 )

3.23 + 0.6 mm (2 - 5)

p = 0.48

Tib-Fib Clear space

Pre-Op

7.04 + 2.1 mm (4 - 12)

7.82 + 1.6 mm (4 - 10)

Post-Op

4.04 + 0.8 mm (2 - 6 )

5.0 + 1.8 mm (3 - 8)

p < 0.05

Tib-Fib Overlap

Pre-Op

3.95 + 2.0 mm (0 - 8)

3.78 + 2.3 mm (0 - 8)

Post-Op

8.21 + 2.0 mm (4 - 11)

7.47 + 2.0 mm (4 - 10)

p = 0.22

Based on radiographic criteria of syndesmosis integrity, 9 patients had syndesmotic malreduction. Only 3 patients with true malreduction on CT scan were correctly diagnosed using radiographic parameters while 6 had a false positive result [Table 3.6].

Table 3.6: Comparison of CT scan and radiographs for diagnosis of syndesmosis malreduction

Malreduction on CT scan

Malreduction on radiographs

Total

No

Yes

No

35

6

41

Yes

2

3

5

Total

37

9

46

The radiographic criteria for syndesmosis malreduction included TFCS > 6 mm or TFOL < 6 mm on standard AP ankle radiographs. On CT scans malreduction was diagnosed if there was > 2 mm difference in the width of syndesmosis as compared with normal side.

3.5. Clinical outcomes

Mean time to full weight bearing was 8 + 1.2 (range 6 - 10) weeks in tightrope group as compared to 9.1 + 1.8 (range 6 - 13) weeks in screw group (p = 0.11) [Fig. 3.5]. Mean American orthopaedic foot and ankle society (AOFAS) hind foot score was 89.56 + 8.6 (95% CI 85.83 - 93.29) in tightrope group and 86.52 + 9.6 (95% CI 82.34 - 90.70) in screw group (p = 0.26). Similarly foot and ankle disability index (FADI) score was 82.42 + 11.2 (95% CI 77.56 - 87.27) in tightrope group and 81.22 + 15.6 (95%CI 74.46 - 87.97) in screw group (p = 0.76) [Table 3.6]. Both functional scores were measured on a scale of 0 - 100 with higher scores associated with better functional outcomes. None of the clinical outcome measures differ significantly between the two groups ( t -test) [Fig. 3.6].

Table 3.6: Clinical outcomes

Tightrope group

Syndesmosis screw group

p value

Time to full weight bearing

8.0 + 1.2 weeks (6 - 10)

9.1 + 1.8 weeks (6 - 13)

p = 0.11

AOFAS Score

89.56 + 8.6 (69 - 100)

86.52 + 9.6 (65 - 100)

p =0.26

(95%CI 85.83 - 93.29)

(95%CI 82.34 - 90.70)

FADI Score

82.42 + 11.2 (58.7 - 97.1)

81.22 + 15.6 (47.1 - 98.1)

p =0.76

(95%CI 77.56 - 87.27)

(95%CI 74.46 - 87.97)

Regression analysis was performed to find any significant correlation between the two groups and the clinical outcome score (AOFAS) while adjusting for potential confounders [Table 3.7]. Type of fixation was not significantly associated with the clinical outcome score. Malreduction of syndesmosis on CT scan was the only variable that reached statistical significance when keeping other variables constant with regression coefficient -12.39; t = - 2.43 and p < 0.05 [Table 3.7].

Table 3.7: Regression analysis to determine the predictors of functional outcome.

AOFAS

Coef.

Std. Err.

t

p > I t I

95% Conf. Interval

Syndesmosis Malreduction

    -12.39

   5.102

-2.43

   0.02

-22.7

-2.09

Fixation techniques

       0.29

   2.855

   0.1

   0.91

-5.47

   6.05

Duration since surgery

      -0.05

   0.176

-0.34

   0.73

-0.41

   0.29

Age

0.008

   0.105

0.08

   0.93

-0.2

   0.22

Constant

90.68

   7.025

12.91

       0

76.49

104.87

AOFAS score is used in this regression analysis as the measure of functional outcome.

Syndesmosis malreduction is the only independent predictor of worse functional outcome score. Regression coefficient of -12.39 indicates that the presence of malreduction in this study resulted in reduction of 12.39 points on the outcome score. Coef: regression coefficient; Std Err: standard error; Conf. Interval: confidence interval.

Chapter No. 4

D ISCUSSION

In this study we compared the accuracy and maintenance of syndesmosis reduction, based on computed tomographic scans and its correlation with the clinical outcomes. This study showed that there was significant difference in the mean width of syndesmosis between operated and normal ankles in screw group as compared with tightrope group. Student's t-test was used to compare the means between operated and normal ankles. The p value for screw group was 0.01 as compared to 0.30 in tightrope group confirming that the results were statistically significant. According to our criteria of malreduction there were 5 cases of malreduction in screw group while none of the tightrope group showed malreduction. There was 21.7% increased risk of syndesmosis being malreduced when treated with screw fixation rather than tightrope technique (p < 0.05, Fisher's exact test). This is in accordance with previous literature regarding syndesmosis screw fixation. The incidence of malreduction of syndesmosis has been reported between 16% and 52% (44, 50, 60, 81) . Weening et al (44) reported 16% of malreduction of syndesmosis in patients treated with syndesmosis screw. The diagnosis of malreduction in that study was based on standard radiographic parameters of syndesmosis integrity and demonstrated a direct relation of malreduction with poor functional outcome scores. As the literature has suggested that the standard radiographic measurements are not accurate (7, 39, 82, 83) and sufficient to diagnose syndesmotic malreduction, several authors has used CT scans for this purpose. Gardner et al (60) has reported 52% of syndesmosis malreduction in there series of 25 patients treated with syndesmosis screw based on CT scans as compared to only 24% using  standard radiographic criteria. This is the highest incidence of malreduction, reported so far in the literature but the validity of the results is limited by the lack of comparison with the uninjured ankle and the lack of clinical correlation. Furthermore they considered the difference of more than 2 mm between anterior and posterior measurement of syndesmosis as significant for the diagnosis of malreduction. This criterion is questionable as Elgafy et al (12) has demonstrated in their study of CT measurements of normal ankle syndesmosis that the mean difference in the anterior and posterior width of syndesmosis was 2 mm. When comparing male and female separately the mean difference was 3 mm for male and 2 mm for female (12) . Our study also showed similar variations in anterior and posterior width of syndesmosis. The mean difference was 1.2 mm (range 0 - 3.3 mm) with wider difference in males than females on normal uninjured side [Table 3.2]. Considering the magnitude of normal variations, Gardner et al probably over estimated the incidence of malreduction in their study. Wikeroy et al (50) reviewed 48 patients treated with syndesmosis screw from an earlier randomised controlled study after 8.4 years and also reported 20.8% incidence of malreduction based on axial CT scan when comparing with normal side. Similar to Wikeroy et al our study showed 21.7% incidence of malreduction in screw group.

Radiographic criteria of syndesmosis integrity as described by Harper et al (38) is routinely used in practice to diagnose syndesmosis diastasis despite several reports questioning the accuracy of those parameters . Our study showed no significant difference between the two groups regarding medial clear space (MCS) (p = 0.48) and tibiofibular overlap (TFOL) (p = 0.22) using t-test . Tibiofibular clear space was although significantly wider in screw group than tightrope group (p < 0.05, t-test ). When radiographic parameters were used to diagnose diastasis there were nine cases of malreduction but did not correlate well with the CT diagnosis. Three out of five of the true malreductions were correctly diagnosed by radiographs while there were six false positive. This also confirms the findings of previous studies (7, 39, 82, 83) .

Although, there was a trend towards better clinical outcomes in tightrope group but when adjusted for potential confounders such as age and duration since surgery there was no statistically significant difference in time to full weight bearing and functional outcome scores (AOFAS, FADI). Malreduction of the syndesmosis was the only independent variable which significantly affected the functional outcome scores.

Similar findings were also reported by weening et al (44) and Wikeroy et al (50) . Accurate reduction of syndesmosis is essential to restore normal biomechanics of ankle joint. Malreduction leads to mismatch in tibial and talar articular surfaces and significantly reduce the contact area and increase the joint reaction forces which can results in early arthrosis and long term morbidity.

Tightrope fixation for syndesmosis injury is a relatively new technique which provides dynamic fixation and obviates the need for routine removal of implant. So far the literature is limited regarding Tightrope and mainly comprises of few case series and nonrandomised comparative studies with limited number of patients and shorter follow up. Thornes et al (71) and Cottom et al (73) compared Tightrope and syndesmosis screw fixation in non randomized comparative study and reported a trend towards better functional outcomes. Thornes et al also performed CT scan in 11 out of 16 patients in Tightrope group after three months and did not find any loss of reduction. None of the syndesmosis screw group had a CT scan limiting the significance of that part of the study. Coetzee et al (76) reported similar trend of better clinical outcomes in there preliminary results of a randomized controlled trail. Earlier studies did not report any complication with this technique but later it has become evident that like any novel technique there is a learning curve and cases of hardware removal has been reported in few studies due to soft tissue irritation over the lateral knot (75-77) . We did not have any complication in Tightrope group requiring hardware removal. In all our cases of tightrope fixation great care was taken to bury the knot deep. We created a periosteal recess at the posterior aspect of fibula before inserting the Tightrope and the knot was buried sub-periosteally. Theoretically, this technique might have helped in reducing the soft tissue irritation over the lateral knot but the association could be just incidental and no hard evidence can be provided on the basis of this study.

So far this is on the only study that compared the accuracy and maintenance of syndesmosis reduction between Tightrope and syndesmosis screw group and showed that Tightrope fixation was significantly better in maintaining the reduction even after a mean duration of 30 months post surgery. The reason for high incidence of malreduction in syndesmosis screw group is hard to determine from this study as CT scans were not performed immediately post operatively which make it hard to discern at which time the diastasis occurred. Whether the syndesmoses were malreduced at the time of surgery or evolved over time. The possibility of increased gap after removal of syndesmosis screw cannot be ruled out. On the other hand Tightrope doesn't require routine removal and thus continue to maintain reduction. As Tightrope is a flexible device one possible explanation of accurate reduction is that fibula is pulled into the concave incisura of distal tibia as it is tightened.

There are several limitations in this study. Firstly, this is a non randomized study and the treatment choice was based on the consultant's preference. As no other variable influenced the choice of fixation, the demographics and the injury pattern in the two groups were comparable. Secondly, it is not possible to identify exactly the reason for higher incidence of malreduction in screw group. It was also not possible to blind the assessor for radiographic and CT measurements as it was obvious which group they belong. To reduce the measurement bias all the measurements were performed by an independent Musculoskeletal Radiologist. CT measurements were repeated at an interval of two weeks in random order, without the knowledge of previous measurements. Intra-class correlation coefficient of 0.91 showed high level of intra observer concordance. Clinical assessment and interviews were performed by an independent assessor who was blinded to the group of patients and two functional outcome scores one clinician reported (AOFAS) and other patient reported (FADI) were used to increase the validity. Despite these limitations, considering the appropriate sample size and follow up duration the results of the study are valid and show that Tightrope fixation is at least equivalent to the conventional screw fixation for the treatment of syndesmosis injuries with potential advantages of providing and maintaining accurate reduction and avoiding need for routine removal. The technique is simple and can be used both in isolation and with plate fixation. It minimise the risk of hardware complication associated with screw fixation and the need for second operation. Like any novel technique, there is a learning curve and care must be taken to avoid soft tissue complications that may require implant removal. Further long term randomized controlled trials would be helpful in clarifying the issue.  

Chapter No. 5

C ONCLUSION

Ankle syndesmosis injuries are complex and require accurate reduction and fixation to restore normal biomechanics of ankle joint and avoid long term complications. Syndesmosis screw and Tightrope fixation are both valid options for the treatment of syndesmosis injuries. Although, short to medium term clinical results were comparable for both the groups, Tightrope provides and maintains more accurate reduction of syndesmosis as compared to screw fixation and obviates the need for routine removal of implant. The radiographic parameters of syndesmosis integrity routinely used are inaccurate and care must be taken to appropriately reduce the syndesmosis before fixation as malreduction of syndesmosis is the most important independent predictor of long term functional outcome.