Feline talocrural luxation: A cadaveric study of repair using ligament prostheses

Summary Currently recommended surgical techniques to treat severe biaxial feline talocrural soft-tissue injuries commonly lead to unsatisfactory outcome. Data relating to canine talocrural stabilisation may not be useful in cats due to major differences in tarsal anatomy between the species. This experimental biomechanical cadaveric study used specimens (n = 10) prepared from the distal pelvic limbs of five adult cats. The aim was to design a technique for treating talocrural luxation using suture prostheses and bone tunnels, and to investigate its suitability for use in clinical cases. Four prosthetic ligaments were placed through a series of five 1.5 mm bone tunnels. Two prostheses, the caudoproximal pair, were taut in talocrural flexion and two prostheses, the craniodistal pair, were taut in extension. The intact specimens had their range-of-motion (ROM) and stability tested, after which they were transected at the talocrural joint (simulated luxation) and repaired using the technique described. The ROM and stability of the repaired specimens were tested and compared to the intact specimens. The repaired specimens had comparable stability to the intact specimens, although the ROM was different (p <0.05) in six of 16 positions (p <0.003125). These corresponded to the positions where the lateral prostheses were taut. The repair technique described may be useful in the treatment of talocrural luxation, as it is lowprofile in an area of limited soft-tissue cover, allows anatomic reduction, restores normal talocrural joint stability and near-normal tarsal ROM.


Introduction
Talocrural luxation results from biaxial ligament injury, and may involve bone fracture, bone loss, and open wounds (1)(2)(3)(4)(5). The subjectively assessed prognosis for cases of severe biaxial feline talocrural softtissue injury is reported to be guarded, although reduction of the duration of postoperative talocrural immobilisation has been suggested to be beneficial (1)(2)(3). Techniques for talocrural repair using suture anchors and prosthetic ligaments have been described for use in feline cases (6,7). However, reports of experimental studies are lacking in this species, though similar studies have been performed in dogs (8)(9)(10)(11)(12). There are major anatomical differences between the canine and feline tarsus, especially regarding the ligaments (13,14). These anatomical differences may make canine repair methods inappropriate for cats. Also, suture anchors may be undesirable in this region because of the minimal softtissue cover.
Treatment aims in cases of joint luxation are to restore anatomical joint alignment, to restore normal joint stability, to maintain a normal range-of-motion, and to allow early joint mobilisation (15)(16). This experimental feline cadaveric study aimed to design a novel repair technique for feline talocrural luxation, using bone tunnels and ligament prostheses, and to test the suitability of this technique by measuring talocrural stability and tarsal range-of-motion before and after talocrural ligament transection and repair.

Anatomical review
The collateral ligaments of the talocrural joint of the dog are described as having long and short parts (8). However the talocrural joint of the domestic cat does not have any long collateral ligaments spanning the whole tarsus, therefore such terms are unhelpful in this species (13,14).
In addition to the ligaments, there are a number of tendons crossing the feline talocrural joint, all of which provide active joint support in the live animal (17). The tendons of the fibularis brevis (lateral) and tibialis caudalis (medial) muscles also provide some passive joint support as they each have a firm attachment to their respective malleolus en route to their insertion distally (13,14). The joint capsule is also complex, and provides passive talocrural joint support in all planes (13,14). Laterally, the passive talocrural joint stabilisers in joint flexion are the fibulotalar ligament gical technique for prosthetic replacement of the ligaments following talocrural luxation.

Pilot study
Three feline cadaveric pelvic limbs were used to evaluate optimal bone tunnel locations and suture configurations for treating simulated talocrural luxation. Reconstruction techniques that were not biaxially symmetrical, and that involved greater distance of the bone tunnels from the talocrural joint, failed to provide sufficient stability.

Specimen preparation
The left and right pelvic limbs of five adult Domestic Shorthaired and Longhaired cats, free from orthopaedic disease and euthanatized for reasons unrelated to this study, were disarticulated at the coxofemoral joint, placed in individual plastic bags, and frozen at minus 80 o C. Before use, they were thawed at 3-5 o C for 24 hours then moved to a laboratory at ambient temperature. The specimens were kept moist with a saline spray throughout the testing procedure, and before measurements a manual warm-up was performed (18,19). Fifty cycles were performed by hand, cycling the tarsus through its full range-of-motion without deliberately forcing motion in any direction. This procedure was repeated before every set of measurements.
The skin was removed down to the phalanges, then the tibia and fibula were osteotomised transversely with a hand saw 100 mm proximal to the medial malleolus, in the region of the distal end of the tibial crest. The soft tissues were also transected at the level of the tibial osteotomy and the proximal portion of the limb was discarded. The muscles and tendons that make up the common calcaneal tendon were removed by cutting the common calcaneal tendon adjacent to the calcaneus and stripping these tendons and muscles away from the remaining specimen. A 2 mm diameter Steinmann pin was drilled parallel to and between the third and fourth metatarsal bones from distal to proximal into the distal bones of the tarsus until it was held firmly (the metatarsal pin). A 2.4 mm diameter 130 mm long Steinmann pin was inserted into the exposed medullary cavity of the tibia until it encountered the distal tibial metaphyseal bone (the tibial pin). This was the intact specimen (IS) (n = 10), and all IS underwent range-of-motion and stability testing as described below. and the fibulocalcaneal ligament (13,14). In extension, only the tendon of the fibularis brevis muscle acts as a lateral passive joint stabiliser. Medially, the caudoproximal part of the tibiotalar ligament provides passive stabilisation to the talocrural joint during flexion. During extension, the passive medial talocrural joint stabilisers are the tibiocentral ligament, the craniodistal part of the tibiotalar ligament, and the tendon of the tibialis caudalis muscle (ǠTable 1, ǠFig. 1) (13,14). Therefore in flexion, the caudoproximal structures (namely the fibulocalcaneal ligament and caudal fibulotalar ligament [lateral] and the caudoproximal part of the tibiotalar ligament [medial]) are taut, and the remaining structures are slack. In extension the craniodistal structures (fibularis brevis tendon [lateral] the tibiocentral ligament, craniodistal part of the tibiotalar ligament, and the tibialis caudalis tendon [medial]) are taut, and the remaining structures are slack. There is a transition between talocrural flexion and extension, where none of the passive joint support structures are taut. The distal tibiofibular ligament connects the tibia and fibula together firmly, keeping the articular surfaces of these bones adjoined so that they function as a single unit.
Based on this review of the functional anatomy of the feline talocrural joint, this study aimed to develop and evaluate a sur- Table 1 Passive stabilisers of the talocrural joint (see also Fig. 1

Transection and repair
All soft-tissue structures crossing the talocrural joint including ligaments, tendons and joint capsule were transected, severing the specimen at the level of the talocrural joint to simulate talocrural luxation. Material was excluded from the study if there was evidence of osteoarthritis in the talocrural joint. The transected specimens were repaired using the following technique: Five 1.5 mm diameter bone tunnels were drilled (ǠTable 2, ǠFig. 4). The calcaneal tunnel was located at the insertion of the fibulocalcaneal ligament on the lateral proximal aspect of the calcaneus, and was drilled lateromedially. The malleolar tunnels were placed in the centre of the most distal part of the medial and lateral malleolus, drilled caudocranially. The distal calcaneal and central tarsal bone tunnels were drilled in the palpable bony prominence located at the distal dorsolateral region of the calcaneus and the dorsomedial aspect of the central tarsal bone respectively; these tunnels were drilled in a dorsoplantar direction and aimed slightly abaxially. Four prosthetic ligaments were placed, using 3.5 metric polydioxanone a , and tied using a sliding knot with five throws (ǠTable 3, ǠFig. 5) so that the paired (lateral and medial) caudoproximal prostheses were taut during full talocrural flexion and the paired (lateral and medial) craniodistal prostheses were taut at full talocrural joint extension*. These were the repaired specimens (RS) (n = 10). Repaired specimens underwent range-of-motion and stability testing as described below.

Range-of-motion testing
The position of the tibial pin relative to the metatarsal pin was measured and recorded, to define the range-of-motion of the tarsus.

Specimen set-up
The specimen was held secure by the metatarsal bones, plantar-side down, in a custom-made gantry so that the metatarsal pin was in the horizontal plane. This pin remained fixed throughout testing. The crus, containing the tibial pin, could move freely within the neutral and elastic zones of the passive joint stabilizers (ǠFig. 2). Table 2 Positions of the bone tunnels (see also Fig. 4).

Table 3
Technique for placement of the prosthetic ligaments. All four prostheses were pre-placed, then the talocrural joint was anatomically reduced. The prostheses were then tied tightly in the order and joint position described.

mm bone tunnel
Specific location
a Polydioxanone, PDS: Ethicon, Kirkton Campus, Livingston, Scotland * Authors' Comment -Anatomic terminology: The terms 'cranial' and 'caudal' are appropriate for use within the crus, down to the talocrural joint, and the terms 'dorsal' and 'plantar' are appropriate for use within the pes, starting from the talocrural joint. Within the pes, 'proximal' and 'distal' refer to the position of structures within that limb segment relative to the body, at the normal standing position. The prostheses have been grouped into lateral and medial pairs of similar function in this paper to echo the functional anatomy of the ligaments in this region. The prostheses cross the talocrural joint, and the pair names 'craniodistal' and 'caudoproximal' reflect the relative position of the origin (on the crus) and insertion (on the pes) of the line of tension of each prosthesis relative to the centre of rotation of the talocrural joint.

Geometric description of articulation and measurement of corresponding angles
The three-dimensional location of the tibial pin relative to the metatarsal pin was defined using a standard spherical coordinate system, in terms of azimuth and elevation angles. These angles are illustrated in ǠFigure 3. The azimuth gives the orientation of the tibial pin projected perpendicularly onto the horizontal plane. The angle was measured in a clockwise direction relative to the position of the metatarsal pin for right pelvic limb specimens (anticlockwise for left pelvic limb specimens), with the metatarsal pin located at zero degrees. The elevation angle measures the tibial pin location above the horizontal plane. With these coordinates, both an azimuth and elevation of zero degrees (physically impossible) would correspond to tarsal flexion with the tibial pin pointing back along the metatarsal, while an azimuth of 180° and an elevation of zero degrees would correspond to full tarsal extension. Azimuths between zero and 180° involve the tibial pin tip being laterally located (i.e. tension in the medial passive joint stabilizers), while azimuths between 180° and 360° correspond to the tibial pin tip being medially located (tension in the lateral passive joint stabilizers). The azimuth was measured using a video camera pointing vertically down towards the anticipated intersection point of the axes of the metatarsal and tibial pins. Articulation data were recorded at azimuth intervals of 22.5° (giving 16 measurements) and a template marked with lines at intervals of 22.5° was placed directly below the specimen to facilitate the measurements. The elevation was measured using a hand-held goniometer fixed in a vertical plane with one arm parallel to the tibial pin and one arm parallel to the projection of the tibial pin on the horizontal plane. This ensured that the true elevation of the tibial pin above the plane containing the metatarsal pin was measured (20). The goniometer and the video image could be seen simultaneously by the experimenter, ensuring that the azimuth and elevation measurements corresponded.

Tibial pin positioning
To measure the range-of-motion, the tibial pin and crus were allowed to rest under their combined weight. No other force was applied in the vertical plane. The azimuth was varied by the experimenter, and the tibial pin was prevented from moving away from the desired azimuth by allowing it to abut against the experimenter's finger-tip. The elevation at each azimuth was determined by the complex interaction of the turning moment of the combined weight of the crus and the pin pivoting on the talocrural joint surface and causing tension in the various passive stabilizers. The magnitude of the tension and the relative position and orientation of these passive joint stabilizers, and the contact area, and angulation of the talocrural joint surfaces, changed with each azimuth position. The elevation as a function of azimuth defined the rangeof-motion, representing the neutral zone plus part of the elastic zone of joint motion.
The same specimen was used for the IS and then the RS, and therefore the crus and pin weight and the talocrural joint topography were identical for these specimens, meaning that the effective gravitational force applied to the passive joint stabilizers was the same for the IS and RS at each azimuth. This force was not quantified due to the complexity of the biomechanics involved, and the extent of the elastic zone involved at each azimuth was also not quantified. The only difference between the IS and RS was in the passive joint stabilizers, which were ligaments in the IS and prostheses in the RS, and so it follows that any difference in the elevation recorded between the IS and RS would be as a result of differences in the relative position and orientation, tension and deformation, of these passive stabilizers.
The use of the experimenter's finger to maintain the azimuth did not influence the Diagram showing the method of calculation of the position of the tibial pin relative to the metatarsal pin, using azimuth and elevation (a standard spherical coordinate system).

© Schattauer 2012
Vet Comp Orthop Traumatol 2/2012 elevation of the tibial pin, only the azimuth, and therefore the magnitude of the force between the tibial pin tip and the experimenter's finger (which was not quantified) is not relevant to the results of this study. The possible influence of friction in the vertical plane between the experimenter's finger-tip and the tibial pin tip was minimized by using clean, dry gloves. The resulting component of force was relatively small, and similar between IS and RS, but was a possible source of variation in the results.

Stability testing
Stability was assessed with the crus and tibial pin resting in the vertical plane and abutting the experimenter's finger in the horizontal plane to maintain its position at each of the sixteen azimuth angles as already described, and in the central neutral position with the tibial pin held vertically. Stability was assessed manually with five degrees of freedom, by holding the midcrus and the distal tarsus and metatarsus and applying force to the talocrural joint, attempting to induce translation of the crus relative to the talus in the following directions: cranial, caudal, lateral, medial. Next, forces of compression then distraction were applied, followed by torsion internally then externally, then combined torsiondistraction testing, internally then externally. The sixth degree of freedom (joint rotation within the vertical plane of the tibial pin) was not tested in this part of the experiment, as the neutral zone was already defined in this plane by the range-ofmotion testing. The forces applied were not quantified, but were small as they were intended to identify any further neutral zone in each position and therefore only needed to overcome friction of the articular surface -it was not the intention to test the strength of the construct or the size of the elastic zone of the passive joint stabilizers. The specimen was simply described (at each position for each force) as stable if there were no palpable laxity or subluxation, or unstable if there were palpable laxity or subluxation. Specifically, subluxation could be detected as a gliding joint movement resulting in focal impingement Diagram showing the drill position and orientation for the five bone tunnels.

Fig. 6
Mean (+ 95% confidence interval) elevation of the intact specimens and repaired specimens at each azimuth. As the azimuth increased, the talocrural joint moved from flexion to extension with the medial joint support structures taut, and back to flexion again with the lateral joint support structures taut.
of a talocrural joint surface edge on the opposite flat joint surface, causing a palpable concentration of forces at this point as well as a widening of the joint and a loss of congruity. Laxity was palpated as a loss of joint surface contact to distraction forces, without the impingement of a joint surface edge on the opposite joint surface. The forces used were similar to those used by the experimenter clinically to assess talocrural joint stability in an anaesthetised cat, but were not quantified.

Statistical testing
Range-of-motion data were tested for normality using the Shapiro-Wilks test. The mean (and 95% confidence interval of the mean) elevation was calculated for the IS and RS at each azimuth using a commercial statistical software package b . Groups compared were left-limb IS to right-limb IS, and all IS to all RS. Range-of-motion measurements were compared using twoway analysis of variance (ANOVA), followed by a series of paired two-tailed t-tests. An alpha value of 0.05 was used, and to allow for the multiple statistical tests used during range-of-motion testing Bon-ferroni correction was used, with a resultant alpha value of 0.05/16 = 0.003125. Stability assessments were compared descriptively.

Results
There was no evidence of osteoarthritis in any specimen. Data were normally distributed.

Anatomical observations during pilot studies
During pilot studies, the talocrural joint which was stripped down to the passive joint stabilizers (see above) was observed to have two separate axes of rotation during flexion and extension. When moving from full talocrural joint extension to flexion, the crus initially pivoted smoothly on the proximal joint surface of the talus (first axis of rotation) with the craniodistal ligaments (see above, ǠTable 1, ǠFig. 1) taut (isometric) during this stage of rotation. Next, the crus slid over the cranial talar body articular surface, with all the ligaments slack, before the caudoproximal ligaments became taut (ǠTable 1, ǠFig. 1). The crus pivoted on the distal part of the curved articular surface of the talar body (second axis of rotation), and the caudal ligaments remained taut (isometric) until the talocrural joint was fully flexed.

Intact specimens
There was no difference in the range-ofmotion of the left and right limbs by ANOVA testing (p >0.05). All left and all right specimens were unstable to distraction and distraction-torsion forces in the neutral position -laxity was palpable with distraction forces applied, and subluxation was palpable with distraction-torsion forces applied. All IS were stable to all other forces in all other positions.

Repaired specimens
All RS were unstable to distraction and distraction-torsion forces in the neutral position -laxity was palpable with distraction forces applied, and subluxation was palpable with distraction-torsion forces applied. All RS were stable to all other forces in all other positions. Subjectively, the stability of the RS was identical to the stability of the IS. The range-of-motion was significantly different to that of the IS by ANOVA (p <0.05). There was significantly higher elevation, indicating a restriction in joint motion, at azimuths between and including 202.5° and 315° when compared to the IS (p <0.003125) (ǠFig. 6). This meant that the passive joint stabilizers on the lateral side of the talocrural joint were taut at a higher elevation in the RS than in the IS, and allowed less talocrural varus whilst moving from flexion to extension and back.

Discussion
This cadaveric study showed that restoration of joint stability that was subjectively assessed as being normal is possible after talocrural joint luxation in the cat, using a bone-tunnel and suture technique designed to mimic the action of the talocrural ligaments. However there was some decrease to talocrural varus during flexion and extension in this model.

Functional anatomy
The motion of the talocrural joint was found to be more complex than expected, with two axes of rotation identified. This finding is of interest when considering talocrural joint repair, as techniques failing to take this into account could lead to suboptimal results. Hinged external skeletal fixation has been used to support the repair of uniaxial talocrural injuries, with a good outcome reported, although tarsal flexionextension was reduced whilst the hinged external skeletal fixation was in place (15). Hinged external skeletal fixation relies upon the precise alignment of the single axis of rotation of the hinge with a single axis of rotation of the joint, allowing uniplanar motion. The presence of two axes of rotation in the normal talocrural joint may have contributed to the reported decrease in range-of-motion (15). None of the talocrural passive joint stabilisers are isometric throughout talocrural flexion and extension. Instead, the tibiocentral ligament and craniodistal part of the tibiotalar ligament are isometric during crural movement about the proximal axis of rotation during late-to-full joint extension, along with the tendons of the tibialis caudalis and fibularis brevis muscles (which function as passive stabilisers due to their strong attachments to the malleolus and the proximal tarsal bones adjacent to the talocrural joint [13,14]). The fibulotalar ligament, fibulocalcaneal ligament and caudoproximal part of the tibiotalar ligament are isometric during crural movement about the distal axis of rotation during late-to-full joint flexion. These observations could be confirmed using a more accurate three-dimensional measuring technique, such as kinematic analysis or by use of a digitizing system, combined with tension measuring device (21).

Experimental protocol
The feline tarsus is a hugely complex joint to attempt to model (17). This study used a new 360 o multiplanar system for measuring joint range-of-motion. The system used by Aron and Purinton involved the calculation of canine joint angles in the three planes of motion, by measurement of angles between pins drilled through the distal crus and the tarsus (8,9). These pins were found to interfere with placement of the ligament prostheses during the pilot study. Kinematic analysis of joint motion was considered, and rejected for a similar reason, as the implant system interfered with the markers. The system used in this study did not aim to allow assessment of the different components of joint motion, namely flexion-extension, valgus-varus, and torsion-rotation, but instead aimed to allow assessment of the net movement of the joint, including (but not quantifying) coupled motion. As described earlier, the differences in the measured data between the IS and RS were solely due to differences in the function of the passive joint stabilizers, and as such this method was suitable for comparing the specimens. Other studies, using cadaveric models to assess repair of the luxated elbow or the subluxating shoulder after medial collateral ligament injury, have used a similar system of measurement of joint angles before and after simulated injury and repair (19,22). However, these studies only made measurements in a single plane, perceived by the authors to be the most important, rather that assessing the impact of the repair on the entire range-ofmotion and stability of the joint, as was performed in this study. The use of more specimens could have allowed detection of smaller differences in range-of-motion, reducing the potential for type 2 errors. Repeated measurements (by one observer or by multiple observers) could have also decreased variation during measurement, such as that induced by friction in the vertical plane between the experimenter's finger and the tibial pin tip as described earlier.
The protocol for assessment of joint stability was similar to that used in other studies, and allowed subjective detection of the neutral zone of joint motion (slack) leading to laxity or subluxation (8-11, 15, 19). The complexity and small size of the feline tarsal joint, and the need for placement of bone tunnels and prosthetic ligaments near to the talocrural joint in this study, and lack of availability, precluded the use of a more objective system as described for the range-of-motion measurements.

Repair technique for talocrural luxation
The medial and lateral malleolar tunnels were placed as close as possible to the origin of the talocrural ligaments, thus more likely optimizing isometry during the different phases of joint motion. This tunnel position also ensured that the prostheses were as short as possible, to limit the magnitude of prosthesis deformation during load, which could have led to construct failure by allowing joint luxation. The proximal calcaneal tunnel was placed at the insertion of the fibulocalcaneal ligament, to allow the lateral caudoproximal prosthesis to mimic the fibulocalcaneal ligament. However, the medial aspect of this tunnel is not situated at the insertion of the tibiotalar ligament as this point is not surgically accessible, so the caudal medial prosthesis additionally spanned the talocalcaneal joint. Tarsal motion was seemingly unaffected by this prosthesis in the current experiment. The distal dorsolateral and distal dorsomedial calcaneal and talar tunnels were placed on the line of the tibiocentral ligament, tibialis caudalis tendon and fibularis brevis tendon, allowing the craniodistal prostheses to become taut during late-to-full extension and hence to mimic these structures. The different tautness of the cranial pair and caudal pair of prostheses in flexion and extension meant that the prosthetic ligaments effectively mimicked the function of the normal ligaments throughout motion. The palpable laxity and subluxation of the RS in the neutral position was similar to that palpated in the IS, and highlighted the functional accuracy of the repair.
It should be noted that this model of talocrural luxation did not involve damage to the distal tibiofibular ligament. If major ligamentous damage is present in the talocrural joint of a cat, it is important to assess the tibiofibular ligament, and stabilize this element of the articulation (23).

Range-of-motion
The RS had a similar range-of-motion to the IS in flexion, extension, and when the medial ligaments or prostheses were taut through joint motion (azimuths 0-180° and 337.5°). An increase in elevation was seen in the RS when the tibial pin was moved from talocrural extension to flexion with the lateral prostheses taut (azimuths 202.5-315°), meaning that talocrural varus was restricted in this phase of joint motion. In clinical cases repaired using this technique, this might restrict proper (plantar) paw placement during joint motion whilst the limb is abducted, resulting in a tendency to walk on the medial aspect of the foot. The impact of this decrease in range-ofmotion on the clinical outcome is not known, but we would expect it to be of minor significance to overall limb and cat function. The polydioxanone suture used will degrade over time, eventually becoming fully absorbed, which could reduce the long-term impact of this decreased rangeof-motion were this technique to be used clinically (24). It is interesting to note that the IS had a dissimilar range-of-motion when the lateral and medial ligaments were taut (azimuths 22.5°-157.5° and 202.5°-337.5°), indicating that paw placement parallel to the ground during talocrural joint motion is possible with the limb abducted to a greater degree than when it is adducted. It is possible that this mild restriction of range-of-motion could have been avoided by tying the lateral prostheses with the joint held in a slightly different position, leading to slightly longer lateral prostheses. However, this could also reduce the stability of the construct and lead to subluxation or luxation; this was not tested and thus cannot be recommended based on this work.

Clinical relevance
Treatment of severe biaxial feline talocrural soft tissue injury, with or without bone loss or fracture, was associated with a moderate or worse outcome in 17/17 cats treated with internal fixation and a three to six week period of talocrural immobilisation postoperatively, whereas treatment outcome of uniaxial feline talocrural injury cases was good or excellent (1)(2)(3). Treatment of biaxial malleolar fracture led to a good to excellent outcome in 13/15 cats (2,3,25). The poor outcome of the biaxial soft-tissue injury cases could be due to the severity of the initial trauma, which may cause these cases to have more cartilage and tendon damage than uniaxial cases. Alternatively, biaxial soft tissue injuries may be technically more difficult to reduce anatomically than uniaxial injuries or bimalleolar fracture cases. Intact stabilisers on one side of the joint in cases with uniaxial injury, or intact tibiotalar, tibiocentral, fibulocalcaneal and fibulotalar ligaments (and intact fibularis brevis and tibialis caudalis tendons) in cases of bimalleolar fracture, may improve anatomical reduction and stability, leading to a better outcome. Non-anatomical reduction, followed by joint immobilisation allowing healing in an imperfect position, may result in altered joint forces on recovery, predisposing to osteoarthritis and pain (26).
The other reason for the reportedly poor outcome of biaxial soft-tissue injury cases could be the three to six week period of transarticular fixation. Four cats with severe biaxial talocrural injury treated with internal fixation and a 10-day period of talocrural immobilisation had a good or excellent outcome (3). Joint immobility provides protection to the injured site during healing, but leads to osteoarthritis if prolonged (27). The limited flexion and extension afforded by hinged external skeletal fixator during early recovery may be beneficial (15).

Future work
The use of our technique could allow early joint motion to be achieved postoperatively, which ultimately may improve the clinical outcome. However, investigation of the most suitable prosthesis type, size, and means of securement is warranted, as is testing of construct strength and durability. The practicality of applying the repair technique to clinical cases is the subject of ongoing investigation, specifically regarding potential complications of prosthesis placement. One change made from the herein reported technique involves the use of two parallel tunnels spaced 1-2 mm apart instead of the single lateromedial calcaneal tunnel, as this has been found to allow easier placement of the caudoproximal prostheses (ǠFig. 7).

Conclusions
The repair technique for feline talocrural luxation described in this study could be useful to help improve the treatment of clinical cases for the following reasons. Firstly, reliable anatomical joint alignment is achieved and more normal joint stability is restored using low-profile degradable implants. Secondly, the resulting tarsal range-of-motion is near-normal. Flexion-extension and talocrural valgus during joint motion was normal in this study, and the clinical impact of the reduction in talocrural varus during joint motion is not clear but warrants further investigation. It is likely that the lateral and medial repair techniques could be useful independently of each other in cases with uni-axial injury, although this was not investigated in this study.