Short and long-term outcome following surgical stabilization of tarsocrural instability in dogs

Summary Objectives: To evaluate the outcome and complications following surgical stabilization of canine tarsocrural luxations. Methods: Medical records of dogs which were surgically treated for tarsocrural joint instability between February 2007 and June 2014 were reviewed. Surgical technique, complications and long-term outcome (via questionnaire and Canine Brief Pain Inventory) were assessed. Results: Twenty-four dogs (26 joints) were included. All injuries were traumatic. All joints had associated fractures; malleolar in 21/26 limbs (13/26 medial). Eight joints had internal fracture fixation and transarticular external skeletal fixator, six had external fixator alone, four had prosthetic ligaments with external fixator, and four had prosthetic ligaments with external coaptation. Two joints had pantarsal arthrodesis and two primary ligament repair. Complications occurred in 24/26 limbs giving 45 distinct complications; 16 were minor, 29 major, and 31 complications were external fixator associated. Prosthetic ligaments were significantly associated with major complications (p = 0.017); five out of eight required subsequent removal between 105–1006 days. Cost was significantly associated with major complications (p = 0.017) and soft tissue wounds (p = 0.03). Long-term lameness was seen in nine of 14 dogs. There was no association between pain severity (p = 0.3) and pain interference scores (p = 0.198) when comparing stabilization methods. Clinical significance: Complications are common; however many are external fixator related. Prosthetic ligaments are significantly associated with major complications. Regardless of technique, a degree of ongoing lameness is likely.

ligaments with external coaptation. Two joints had pantarsal arthrodesis and two primary ligament repair. Complications occurred in 24/26 limbs giving 45 distinct complications; 16 were minor, 29 major, and 31 complications were external fixator associated. Prosthetic ligaments were significantly associated with major complications (p = 0.017); five out of eight required subsequent removal between 105-1006 days. Cost was significantly associated with major complications (p = 0.017) and soft tissue wounds (p = 0.03). Long-term lameness was seen in nine of 14 dogs. There was no association between pain severity (p = 0.3) and pain interference scores (p = 0.198) when comparing stabilization methods. Clinical significance: Complications are common; however many are external fixator related. Prosthetic ligaments are significantly associated with major complications. Regardless of technique, a degree of ongoing lameness is likely.

Introduction
The tarsocrural joint is formed by the tibia, fibula, talus and the calcaneus (1). Tarsocrural instability is an uncommon distal limb injury in dogs generally involving fractures of one or more of the bones con-tributing to the joint, varying degrees of ligament impairment, or a combination of both (1)(2)(3)(4)(5). The tarsocrural joint is particularly prone to fractures and shear injuries due to the paucity of soft tissue protection in this area (2,6,7). Injuries commonly occur following road traffic accidents, re-sulting in skin, muscle, ligament, and bone injury (1,2,7).
The anatomy of the tarsus is complex, often making diagnosis and management challenging (1,3,8,9). Initial management of tarsocrural instability aims to limit further damage to the articular surface and supporting soft tissue structures, allowing restoration of anatomical joint alignment with stability to facilitate healing (3,10,11). Treatment modalities include combinations of primary ligamentous repair, prosthetic ligament reconstruction, external coaptation, transarticular external skeletal fixation, arthrodesis, and amputation (1-5, 7, 10, 12-15). Management with external coaptation alone can be inconvenient, poorly tolerated, and may result in coaptation associated soft tissue injuries (16). Some injuries of the tarsocrural joint are too extensive to be successfully reconstructed, leading to arthrodesis in order to maintain limb function (1,2). Arthrodesis with a plate or external fixator may also be used as a salvage procedure if other methods of stabilization have failed (1,2,7).
To date, no studies compare treatment outcomes following surgical stabilization of tarsocrural joint instability in dogs. The purpose of this study was to retrospectively evaluate the outcome and complications following surgical stabilization of canine tarsocrural luxation/subluxations. In addition, the study aimed to evaluate differences in functional outcome as assessed by owner questionnaire.

Material and methods
Medical records of dogs with tarsocrural joint instability treated surgically between February 2007 and June 2014 were re-viewed. Tarsocrural instability was defined as palpable instability at that joint level, then confirmed as loss of articulation between the talus and the tibial cochlea on survey or stressed radiographs (▶ Figure  1). The following information was gathered for each patient: signalment, injury, cause of injury, concurrent fractures, presence of soft tissue wounds, duration of hospitalization, number of follow-up examination visits, complications, cost of treatment, and stabilization method.
The luxations and fractures were stabilized with internal fixation when appropriate. Primary collateral ligament repair was attempted if possible when instability was attributable to ligament damage. Complications were categorized as minor or major. Minor were defined as those not requiring additional surgical treatment. Major were those requiring further surgical treatment. Soft tissue wounds were divided into minor or major. Minor included superficial abrasions and puncture wounds. Major included all wounds other than superficial abrasions and puncture wounds.
Final outcome of each dog was assessed by owner questionnaire consisting of two sections: section one assessed owner satisfaction, ongoing medication and long-term complications. In section two, owners assessed long-term function and pain using a validated client questionnaire; the Canine Brief Pain Inventory (CBPI) (17,18). The CBPI assesses owner perception of pain severity and pain interference. The pain severity questions were scored on a scale of 0 (no pain) to 10 (extreme pain). The pain interference questions i.e. how much pain interfered with the dog's normal function, were scored on a scale of 0 (no interference) to 10 (completely interferes). The responses to these questions were averaged to generate the pain severity and pain interference scores (17,18).
Commercially available statistical software programmes a,b were used to perform all statistical analyses. Data were assessed for normality using the Shapiro-Wilk test. Associations between the presence of wounds, major complications, minor complications, fractures, non-tarsal injuries, soft tissue injuries, presence of an external skeletal fixator, and the final cost of treatment were assessed using the Mann-Whitney U test. The same associations were assessed in relation to hospitalization time. Fisher's exact test was used to determine associations between the stabilization type and presence of complications. The Kruskal-Wallis test was used to compare pain severity and pain interference scores between treatment groups. Treatment groups were allocated as shown in ▶ Table 1. Pain interference and severity score association with talar fractures and wounds was as-  Eight joints had internal fracture fixation and a transarticular external skeletal sessed with the Mann-Whitney U test. Association of weight and complication development was assessed using t-test. A p <0.05 was considered significant.

Results
Twenty-four dogs with surgically managed tarsocrural joint instability met the inclusion criteria. Age on presentation ranged from 10 months to 10 years 10 months (median: 4 years 11 months), and weight ranged from 10 kg to 43 kg (mean: 27 kg). Breed and sex distribution are outlined in ▶ Appendix 1 (Available online at www.vcot-online.com). All recorded injuries were traumatic in origin; 13/24 dogs sustaining a road traffic accident, four out of 24 developed an injury whilst running and three out of 24 fell from a height. The remaining known causes included being trodden on and limb entrapment. Suspected trauma was reported in one

A B
Eight limbs were stabilized with prosthetic ligaments; six using multifilament fiberwire and two with monofilament nylonleader-line. The two joints stabilized using nylon and three of the six joints stabilized using fiberwire developed major long-term complications due to infection and required implant removal. Four of five infected prostheses also developed clinical joint instability, which was absent prior to implant infection. Owners reported swelling or sinus tract formation from 105-1006 days postoperatively (median: 156 days) (▶ Figure 3). Placement of prosthetic ligaments was significantly associated with postoperative complications (p = 0.017) compared to limbs which had no prosthetic ligaments placed when fixator-associated complications were excluded. Two dogs, (numbers 1 and 19) required revision surgery following implant removal after prosthetic ligament infection. Both dogs had prosthetic ligament removal and subsequent stabilization with a transarticular external skeletal fixator in dog 1 and plated pantarsal arthrodesis in dog 19 (▶ Figure  3). The two dogs (numbers 7 and 10) with no complications were stabilized by both fiberwire prosthetic ligaments combined with postoperative coaptation and plated pantarsal arthrodesis respectively. Total cost of referral treatment was significantly increased if major complications occurred (p = 0.017), or tarsal soft tissue wounds were present (p = 0.03). No significant association was seen between cost of treatment and development of minor complications, the presence of non-tarsal soft tissue injuries, fixator placement or the presence of non-tarsal fractures. Similarly the development of major or minor complications and soft tissue injuries had no significant association with hospitalization time. Patient weight was not associated with development of minor (p = 0.86), major (p = 0.27), or non-fixator related complications (p = 0.73). Canine Brief Pain Inventory Mean postoperative pain severity scores and pain interference scores are shown for all patients with available CBPI in ▶ Appendix 2 (Available online at www.vcot-online.com). No significant association between pain severity (p = 0.3) or pain interference score (p = 0.198) were identified when comparing surgical stabilization techniques. No significant association between pain severity (p = 0.164) or pain interference (p = 0.77) score was identified when comparing dogs with and without talar fractures. Similarly, no association was seen when comparing dogs with and without major wounds, (p = 0.494) and (p = 0.29) respectively.

Discussion
Canine tarsocrural instability leads to severe loss of limb function. All our patients were managed surgically, whereas in hu-

A B
mans the question of surgery versus conservative treatment for ankle fractures remains controversial and is influenced by the specific injury combination (19,20). The difference in approach between human and veterinary patients may in part lie in the plantigrade nature of the human pes with its inherent mediolateral stability, whereas the canine digitigrade stance continually loads the tarsocrural support structures in the stance phase. Human patients are also more amenable to resting for extended periods. There are several surgical stabilization techniques available, however assessment of long-term outcome and surgical complication rate were not previously available in dogs. Whether human or veterinary, treatment aims are to reestablish anatomical reduction of the talus in the ankle mortise and maintain joint stability (19). Generally, results following reduction of human ankle fractures appear to be good, although post-traumatic arthritis has been described in 10% of patients despite anatomical reduction (21). This study showed that there is generally a reasonable outcome following a variety of surgical techniques in canine patients; however a degree of permanent lameness is expected regardless of fixation type, and minor complications are very common. In this study, several surgical methods of stabilization were used; however all included tarsocrural joint reduction with immobilization. Many had reduction and immobilization alone using a transarticular external skeletal fixator. Fixator application alone is well documented in canine shear injuries, in one study, six out of seven canine distal limb shear injuries were stabilized with a transarticular fixator (7). The aim of joint stabilization is to provide sufficient support until the periarticular tissues including the ligaments and joint capsule, can heal and fibrose sufficiently to provide stability. We found that clinical results from transarticular fixator stabilization alone were similar to ligament repair or prosthetic ligament placement in addition to temporary immobilization. A small number went straight to salvage with pantarsal arthrodesis, which has previously been advocated for salvage of severe tarsal injuries in both dogs and cats (5,22). Interestingly, regardless of the method chosen, long-term outcome was similar, with a large portion of dogs suffering postoperative complications and long-term lameness.
Transarticular external skeletal fixators are a well-documented joint immobilization technique (2,7,10,11,13,23,24). We found that fixators were used extensively in these injuries as either a sole-fixation device or as adjunct immobilization to protect a primary repair. The fixator complication rate was 19/19 in the present study with pin tract infection in 11/19 and implant failure in 14/19 limbs. Previous transarticular fixator studies reported variable rates of complications ranging from 14% to 71% (7,10). The current study findings indicate a higher overall chance of complications; however this could relate to recording differences, or perhaps this location is particularly vulnerable due to the high loads placed upon a joint-spanning frame. The alternative method of immobilization was coaptation, which can save on intra-operative surgical time and hence cost, although continued dressing changes with coaptation should be considered (23). Following exclusion of fixator-associated complications when comparing fixation groups in the six dogs that had fixator placement as their only stabilization, there was a significant reduction in non-fixatorassociated complications. This probably reflects the fact that 50% of non-fixator complications were coaptation related. Overall, the immobilization method appears to affect the complications seen. We suggest that fixators may still be preferable as although the complication rate was high, they are generally manageable and selflimiting following frame removal, and although in our small coaptation group no major complications developed, it has been previously documented that coaptation has a 63% risk of causing soft tissue injuries (16). The ultimate choice of immobilization however should be based on clinical experience on an individual patient basis.
Some dogs had prosthetic ligament placement in addition to tarsocrural reduction and immobilization. Prosthetic ligaments have been described for medial and lateral collateral ligament replacement in dogs and cats, and they can be an effective way of maintaining range of motion while providing stabilization (3,12,15,25). Use of prosthetic ligaments however was significantly associated with severe long-term complications, occurring up to two and a half years following placement. Previous studies have shown the high potential for complications with up to 50% infection rate with braided material and their use has been advised with caution (2,13). In the current study, monofilament prostheses also required removal due to infection. Importantly, increasing antimicrobial resistance in small animals in conjunction with the increased cost of treatment associated with surgical site infections makes prosthetic ligament use questionable given the high rate of infection and comparable clinical outcome when they are not used (26,27). Therefore, given the increased risk of complications with prosthetic ligaments, the authors would suggest using them with extreme caution, and warn owners of the potential for late complications developing. Four of five dogs that developed implant infection also developed clinical joint instability, which was not present prior to infection and may indicate that prosthesis use reduces long-term periarticular fibrosis development, or infection development can subsequently reduce soft tissue stability. Overall, the emphasis should be on strict aseptic technique if prosthetic ligaments are used.
Primary repair of ruptured collateral ligaments is also possible and can result in excellent outcomes. There were no complications directly attributable to ligament repair, and again, the clinical outcome was not hindered by their usage. Therefore, if the injuries present allow, attempting direct ligament suture repair remains an option, although a monofilament absorbable suture is recommended. This repair would duly need to be protected by either a fixator or coaptation. Plated pantarsal arthrodesis was performed in two dogs as the primary treatment and in another following initial stabilization failure. Previous studies documenting outcome following pantarsal arthrodesis in 40 dogs showed an overall complication rate of 75% with a major complication rate of 32.5%. A high proportion of these were soft tissue related, including catastrophic plantar necrosis associated with injury to the dorsal pedal or perforat-ing metatarsal arteries. A minor complication rate of 42.5% was shown, frequently caused by prolonged external coaptation (4). In the current study, owner questionnaire results are only available for two dogs following plate pantarsal arthrodesis and one dog following primary ligament repair, thus making group size too small to infer substantial conclusions.
Long-term owner questionnaire results showed that nine of 15 dogs were active to very active following surgery with 11/15 owners rating surgical success as good to excellent. Owner satisfaction with treatment was similarly high compared to those reported previously for both tarsocrural instability and shearing injuries (2,7,11,28). Overall owner satisfaction with surgery was high regardless of surgical stabilization type and in spite of the high complication rate. Nonetheless, nine of 15 dogs were reported as having long-term lameness or stiffness by their owners with half receiving long-term non-steroidal anti-inflammatory drugs. This discrepancy could reflect owner counselling at the outset of treatment as to the severity of the injury sustained and the possibility of complications. The alternative to stabilization in many of these cases would be arthrodesis or amputation, and whilst many dogs showed signs of ongoing lameness post stabilization, this may be felt to be a relatively good outcome compared with the alternatives. In other studies, osteoarthritis was documented in 81% of canine joints evaluated following tarsal shear injury, with 23.5% of canine patients suffering long-term lameness (2). Additionally, periarticular fibrosis and post-traumatic osteoarthritis are likely to be a cause of ongoing lameness.
Five dogs with prosthetic ligaments required implant removal; however only two of these owners completed the questionnaire. Overall success of surgery was rated as excellent following revision surgery in one dog which had a subsequent pantarsal arthrodesis using a transarticular fixator. The other dog had its prosthetic ligament removed and plated pantarsal arthrodesis performed, but further longterm implant associated infection was ongoing at the time of data collection (▶ Figure 2) and surgical success was rated as poor (19). Overall, there is an indication that the long-term outcome following implant infection may be guarded, and while dogs can recover, the risk of ongoing infection should be considered prior to choosing any revision stabilization method. Only dog 15 had a long-term pain and severity score of zero. This excellent outcome could be attributable to the nature of the injury sustained and does indicate that full return to function can be achieved.
Inherent limitations of this study include its retrospective nature, use of owner questionnaire and lack of objective gait analysis. Multiple surgeons contributed cases over the study period, which inevitably creates variation in case management and record-keeping. A variety of injury combinations were seen resulting in tarsocrural injury, however due to the small numbers, further stratification was not possible. Multivariate analyses were not performed due to the small number of dogs in the study, and any benefit with regard to surgical technique requires further prospective studies, including objective force plate analysis with increased case numbers.
Owing to our small sample sizes, our statistical analyses were inherently at risk of type two statistical error.
Our study findings confirmed that tarsocrural fractures and luxations are complex injuries to manage. Temporary joint immobilization is essential, and can be successfully used alone or in combination with direct ligament repair or internal fixation of fractures or both as appropriate. Transarticular external skeletal fixators remain the authors' preferred method of immobilization, however fixator complications are guaranteed. Placement of prosthetic ligaments is significantly associated with infection-related complications that typically require further surgery to extract the prosthesis, and these problems can occur over a protracted time frame. The authors therefore would counsel against using prosthetic ligaments as part of the surgical management. Whatever the method of fixation, owner satisfaction appears high, the clinical outcome is reasonable, but a degree of ongoing lameness appears likely. We suggest that owner education is paramount, as expectations for full return to normal function must be managed due to the low proportion of dogs returned to pre-injury status. Given the large case variability, ultimate choice of stabilization must be made on a patient by patient basis with consideration to our findings.