Impact of fixation method on postoperative complication rates following surgical stabilization of diaphyseal tibial fractures in cats

Summary Objectives: To compare the complication rate between open reduction and internal fixation (ORIF) and external skeletal fixation (ESF) for feline diaphyseal tibial fractures. Methods: In a retrospective study spanning a 10 year period, 57 feline tibial fractures stabilized via ESF or ORIF were included for analysis and complication rates were compared between the two methods. Results: In the overall study population, 23 (40.4%) cases suffered complications (9 major, 20 minor, 6 with both major and minor). All of the major complications occurred in the ESF group. Complications were more common in cats with ESF (50.0%) while only one (7.7%) of the ORIF cases suffered complications (OR 12.0 [CI: 2.09; 228.10], p = 0.02). Use of postoperative antibiotic medications was identified as a confounder. After adjusting for confounding, stabilization using ESF remained associated with a higher risk of complications (OR = 13.71 [CI: 2.18; 274.25], p = 0.02). Cats with ESF had a longer duration of follow-up (15.6 weeks; 95% CI: 13.0; 18.3) compared to ORIF (9.5 weeks; 95% CI: 6.4; 12.7) (p = 0.003), and a higher number of revisits (mean 3.0; 95% CI: 2.4; 3.6) than the ORIF group (mean 1.6; 95% CI: 0.9; 2.3) (p = 0.002). Clinical significance: This study demonstrates a significant difference in complication rates between the methods of stabilization, with ESF resulting in a significantly higher complication rate compared to ORIF. Based on these results, it may be prudent to select ORIF for stabilization of feline tibial fractures wherever practical.


Introduction
Tibial fractures occur commonly in cats with the tibia being the second most commonly fractured long bone (1)(2)(3)(4). Tibial fractures are prone to fracture healing complications, including development of delayed and non-union; the tibiae represented 61.1% of feline long-bone fractures developing non-union in one study (1,2,5,6). Multiple factors have been proposed to result in delayed and non-union including inadequate blood supply from the limited extraosseous soft tissues, insufficiently stable fixation, excessive fracture gaps, and interposition of soft tissue in the fracture gap (1,2,(7)(8)(9). Open tibial fractures and comminuted tibial fractures may be more likely to develop non-union (2).
Many different stabilization methods have been described for feline tibial fractures including open reduction and internal fixation (ORIF), external skeletal fixation (ESF), low-stiffness ESF, circular ESF, and minimally invasive plate osteosynthesis (5,(10)(11)(12)(13)(14). Open reduction and internal fixation remains common practice but concerns exist regarding the additional surgical trauma involved with this technique, which may compromise the blood supply to the fracture site hereby delaying bone healing and increasing complication rates (2,5,10,(15)(16)(17)(18)(19)(20). Due to these concerns, minimally invasive plate osteosynthesis has been proposed, which prioritizes soft tissue preservation over absolute mechanical stability (13). External skeletal fixation is frequently recommended for primary or sole stabilization of fractures of the tibia, but several complications are associated with this form of fixation, including pin loosening, pin track discharge, infection, fixation failure, and quadriceps contracture in addition to delayed and non-union (21)(22)(23)(24)(25)(26). External skeletal fixation can also be more intensive in terms of postoperative care for the owner (12). Additionally, there is some debate as to the most appropriate frame configuration for a given fracture. Whilst the use of more rigid ESF configurations has been advocated to enhance fracture healing and decrease complications, two studies have demonstrated that compli-cations occur more commonly following stabilization of tibial fractures using type 2 ESF frames when compared to type 1 frames, and higher complication rates were not noted in one case series reporting the use of low-stiffness ESF for treatment of canine and feline tibial fractures (2,10,11,(27)(28)(29)(30).
Whilst many individual studies have evaluated specific techniques, from these results it remains difficult to compare the available techniques for stabilization of tibial fractures in terms of their associated complication rates. To the authors' knowledge, no study has been performed to compare complication rates following feline tibial fracture repair with ORIF using plates and screws and ESF. The aim of this retrospective study was to review the case records of cats with tibial fractures to investigate whether either ORIF or ESF was associated with a higher rate of complications. Potential confounders included breed, sex, age, presence of other fractures or comorbidities, fracture location and orientation, presence of comminution, fracture type (open or closed), and use of postoperative and perioperative antibiotic medications. Our hypothesis was that ORIF would be associated with a lower postoperative complication rate than ESF for stabilization of feline tibial fractures.

Materials and methods
Medical records at a the Queen Mother Hospital for Animals were searched to identify all cats admitted during a ten year period (January 2004 to January 2014) with an extra-articular, diaphyseal tibial fracture stabilized using either ORIF or ESF. Cats were included in the study if the medical record was complete and radiographs of the fracture site were available for review.
Information collected from the records included signalment (breed, sex, age), limb affected (left, right), fracture location (proximal, mid or distal diaphyseal), orientation (transverse, spiral, oblique or comminuted), and fracture type (open or closed). Open fractures were further classified into grades 1-3 on the basis of soft tissue disruption as has been previously de-scribed (31). The presence of any other fracture affecting the same limb or elsewhere was noted as were any comorbidities. The method of stabilization was recorded as either ORIF or ESF. The frame configuration was also noted where ESF was the method of stabilization (Type 1, 2, 3 or hybrid). The use of perioperative antibiotic medications (classified as up to 24 hours postoperatively) and postoperative antibiotic medications (classified as any use beyond 24 hours postoperatively) were recorded.

Perioperative management
Postoperative analgesia consisted of administration of methadone a (0.1-0.2 mg/kg IV q4-6h PRN) for the first 24 hours, followed by buprenorphine b (0.02 mg/kg IV q6-8h PRN) for the following 48 hours. Meloxicam c (0.1-0.2 mg/kg IV) was administered at induction of anaesthesia unless there were systemic contraindications and it was also prescribed for the following seven days (0.05 mg/kg PO q24h). A selfadhesive wound dressing d was applied to the surgical site until the patient was discharged. Cage rest with gradually increasing duration of supervised periods out of the cage was advised until clinical and radiographic union were confirmed.
The presence of postoperative complications was recorded and the specific details of the complication noted. Complications were defined as any undesirable outcome associated with the surgical procedure and were classified depending on severity. Complications were defined as major (surgical intervention performed) or minor (managed without surgical intervention). The presence of postoperative infection was specifically recorded. As has been previously described, a wound or pin track was considered infected when there was a purulent discharge, abscess or fistula, or when one or more of the clinical signs of pain, localized swelling, redness, heat, fever or deep incision spontaneous dehiscence was identified on physical examination, or when an organism was isolated from an aseptically collected sample by culture or a positive cytology study, or there was some combination of these abnormalities (32,33). The final outcome was classified as success or failure. The case was considered a success if clinical and radiographic signs of fracture union were achieved and the case was discharged with no further visits recommended. For ESF cases, this coincided with the time when frame removal occurred. The case was considered a failure if clinical union was not achieved, necessitating salvage surgery or euthanasia. The number of revisit appointments and the length of time required to reach this outcome were also recorded.

Statistical analysis
Continuous data were expressed as median values and ranges; categorical data were expressed as proportions with 95% confidence intervals (95% CI). The association between fixation method and the risk of any complication (minor or major) was estimated using univariable logistic regression analysis. Univariable logistic regression was used to explore the relationship between potential confounding factors and the risk of any complication, and causal diagrams were used to identify any a priori confounders. Age, a continuous variable, was examined as continuous, categorical, and as a multiple fractional polynomial to determine if there was any evidence of non-linearity in the association with any complication (34). Variables with a p-value of <0.2 were included in the multivariable logistic regression model to explore the presence of confounding (a 10% change in the odds ratio [OR] for fixation method) and test biologically meaningful interactions (35). The fit of the final model was evaluated with the Hosmer-Lemeshow goodness-of-fit test.
To investigate whether ESF patients on average had a longer duration of follow-up and more revisit appointments on average than ORIF patients, a t-test was performed with the group mean and 95% CI compared. The Fisher's exact test was used to explore the relationship between fixation method and postoperative infections. Stat-

Association between fixation method and postoperative complications
All of the major complications occurred in the ESF group. The numbers of complications which occurred in each group are shown in ▶ Table 1. Complications (any type) were more common in cats with ESF (50.0%) while only one (7.7%) of the ORIF cases suffered complications (OR 12.0 [CI: 2.09; 228.10], p = 0.02). All five Siamese cats suffered complications, three of which were major. There were no postoperative infections in the ORIF group, however there were 12 infections in the ESF group (p = 0.05). It was not possible to calculate an odds ratio for this result due to the lack of any infections in the ORIF group.
Four of the variables evaluated met the criteria for potential confounders for the relationship between fixation method and any complications (▶ Table 2) with p-values <0.2 on univariable analysis: presence of fracture comminution, presence of open fracture, sex, and use of postoperative antibiotic medications. No a priori confounders were found. For the purposes of statistical analysis, open fractures were only classified as open or closed rather than taking the grading (1-3) into account. This was because the complication rate was 100% in both the grade 2 and 3 cases which would have made the statistical model less stable. Limb (left or right) was not included despite having a p-value of 0.06 (OR 3.2) because there was no plausible biological reason for this. Following backwards elimination procedure, postoperative antibiotic medication remained as a confounder and no interaction was identified. After adjusting for confounding, stabilization using ESF was associated with a higher risk of complications (OR = 13.71 [CI: 2.18; 274.25], p = 0.02). There was no indication were a type 1 (two with an intramedullary pin tied in), 23 were a type 2, three were a type 3, and three were hybrid. All cases were treated with perioperative antibiotic medications. Twenty-one cases were treated with postoperative antibiotic medications. Postoperative infection, either of the surgical site or of a pin tract, was diagnosed in 12 cases.
Twenty-three (40.4%) cases suffered complications. Of these, nine cases suffered major complications and 20 cases suffered minor complications. Six cases suffered both major and minor complications, giving a total number of 29 complications in 23 cases. Complications encountered included pin track infection and loosening (n = 9), delayed union (n = 5), non-union (n = 8), pin loosening in the absence of infection (n = 1), Kirschner wire migration (n = 1), osteomyelitis (n = 1), mild distal tibial valgus deformity (n = 1), frame disruption leading to fragment displacement (n = 2), and severe lethargy of unknown cause necessitating euthanasia (n = 1). Of the non-union cases, four had revision surgery and four limbs were amputated.
In terms of eventual outcome, one case from the ESF group was lost to follow-up prior to reaching clinical union. There were five failures, (four amputations and one euthanasia) all of which were in the ESF group. The remaining 51 cases were considered successful (including 38 ESF cases and 13 ORIF cases). The number of revisit appointments for each cat ranged from one to 10 (median 2). The length of follow-up ranged from three to 36 weeks (median 12).
istical analyses were performed using three different software programs e,f,g .

Results
A total of 59 cats with tibial diaphyseal fractures that were surgically stabilized with either ORIF or ESF were identified between 2004 and 2014. Two cases were lost to follow-up and were excluded, leaving 57 cats. Of these there were 40 Domestic Shorthair, two Domestic Longhair, five Siamese, two Burmese, and one each of British Shorthair, Bengal, Ragdoll, European Shorthair, Ocicat, Burmilla, Birman and Abyssinian. For statistical analysis these were grouped into Domestic Shorthair / Longhair (n = 42), Siamese (n = 5), and other pedigree (n = 10). There were 37 (35 neutered, 2 entire) male cats and 20 female (16 neutered, 4 entire). The cats were aged between three and 200 months with the median age at 33 months.
Six cats suffered fractures to other bones on the ipsilateral limb, five cats suffered fractures affecting other limbs and four cats had other comorbidities including one cat with pulmonary contusions, two with chronic kidney disease, and one with a peritoneopericardial diaphragmatic hernia.
Out of the 57 fractures, 44 fractures were stabilized using ESF and 13 using ORIF. Of the ESF configurations used, 15 External skeletal fixator frame type was also associated with complications. As ESF type increased by one the OR increased by 2.46 (p = 0.006). This variable could not be added to the multivariable logistic regression model as it was strongly correlated, as would be expected, with the main exposure variable (stabilization using ESF) and thus would cause problems of collinearity.

Discussion
Results of this study demonstrated that in cats with diaphyseal fractures of the tibia, stabilization using ESF resulted in a higher complication rate than stabilization using ORIF. The complication rate for the ESF    (11). Premature pin loosening associated with infection was responsible for 41% of the complications in this study and this was also a frequent contributor in other studies (10)(11)(12). The nonunion rate in this study of 14% was also similar to that reported in previous studies for feline tibial fractures (2,36). Whilst Nolte and others reported an overall nonunion rate of appendicular fractures in cats of 4.3%, when studying tibial fractures in isolation, a 15.3% non-union rate was reported (2).
As has been noted in previous studies, our data also indicated an increasing complication rate as the complexity of the ESF frame increased (2,11). Previous studies investigating fractures of the appendicular skeleton in cats have reported that the use of large and rigid ESF frames may predispose to development of complications, specifically non-union and the high nonunion rate in the ESF group in this study would support this conclusion (2,11). Two potential reasons have previously been put forward to explain this. It may be that the constructs used were excessively rigid thereby decreasing the load borne by the bone and resulting in bone resorption, or it may be that these more rigid constructs were more likely to be used to stabilize fractures which were already predisposed to non-union (2,11). The same explanations may apply to our finding of increased complications for fractures stabilized by increasingly complex frames. However, our results do not support these arguments as explanations of the reduced complication rate for ORIF cases. It could be argued that the higher complication rate in the ESF group for this study could be due to the nature of the fractures stabilized using ESF, for example open fractures are more likely to be stabilized using ESF, and cats with open fractures have been shown previously to be more likely to develop non-union (2). Our statistical model does not support this however as while presence of fracture comminution and presence of open fracture did qualify as potential confounders for this result after univariable analysis, following backwards elimination these were eliminated as confounding factors.
The previously postulated factors leading to the high complication rates for feline tibial fractures have included inadequate blood supply, insufficiently stable fixation, excessive fracture gaps and interposition of soft tissue in the fracture gap (1,2,(7)(8)(9). While inadequate blood supply from the extraosseous tissue may be contributory, it would seem logical that the cases stabilized with ESF would better preserve this blood supply as a large open approach is avoided and therefore one would expect a lower complication rate for ESF cases if this were the predominant factor. Fractures stabilized using ORIF are accurately reduced before internal fixation is applied where possible, and therefore excessive fracture gaps and interposition of soft tissue in the fracture gap would be more likely to be avoided, therefore these may be contributory factors to the reduced complication rate for ORIF cases in this study. However, when minimally invasive plate osteosynthesis stabilization is employed, accurate reduction and clearing of the fracture site are not prioritized prior to plate application and yet, the small number of feline cases reported (eight in one study and four in another [13,14]) have similarly positive results to the ORIF cases in this study, with no major complications and all cases reaching clinical union, and therefore the importance of this can be contested. The argument regarding insufficiently stable fixation is also not clear as if the reduced complication rate in the ORIF cases is due to more stable fixation, it becomes difficult to explain the increasing complication rate in the ESF group as the frame complexity and stability increases. It could be that the ESF constructs are too rigid as mentioned above, and this may suggest a role for early destabilization of these constructs, as has been shown to be beneficial in dogs, but further investigation would be required to assess this (37).
Another finding of significance in this study was the increased length of follow-up and the increased number of revisit appointments required for cats where ESF stabilization was employed. To the authors' knowledge, this has not been reported previously. On average, ESF cases required assessment for six weeks longer and attended double the number of revisit appointments. This was not our primary question to be answered for this study and has therefore not been analysed in as much detail, but it may be useful for managing owner expectations and providing accurate estimates for treatment. Additional appointments, in addition to the higher complication rate associated with ESF stabilization, could add greater expense to treatment costs.
It is interesting that in our case series, the complication rate was 100% for the five Siamese cats. These cases were all treated with ESF, (four with type 2 frames and one with a type 3) and the numbers are too small to be analysed statistically, but it is a finding that potentially warrants further investigation. Three of the non-unions in this study occurred in these five cats, two of which had a successful outcome after revision surgery and one of which had an amputation. The authors are not aware of any reported predisposition of Siamese cats to development of non-union but investigation of this further in the future with larger case numbers is indicated.
In our statistical analysis, the use of postoperative antibiotic medication remained as a confounder following backwards elimination procedure. This was not the primary question that the model was developed to answer, but it remains an interesting finding. The use of postoperative antibiotic medication was at the discretion of the surgeon and was probably related to cases where a break in asepsis occurred, where the fracture was open, or where gross contamination was evident and therefore this finding is not entirely unexpected. However, it has also been demonstrated that in cases where extension of perioperative antibiotic medication is not strictly warranted, extension of administration into the postoperative period results in increased rates of infection (38). It is interesting that in this study, there was no crude association between the use of postoperative antibiotic medications and postoperative infection. This, and the persistence of postoperative antibiotic medication as a confounder following backwards elimination, serve as reminders to clinicians to carefully consider the individual case specifics when deciding whether postoperative antibiotic medication should be continued.
There are a number of important limitations to our study design. The primary aim of the study was to evaluate the association between fixation method and complication rate. As such, the statistical model used was designed to investigate this and this limits the conclusions which can be drawn regarding the other variables mentioned. However, these limitations relate to the secondary findings rather than the primary aim of the study and are therefore considered to be of less importance.
The retrospective nature of the study introduces numerous potential sources of error, particularly with regard to the potential for reporting inaccuracies and the lack of standardization in case management. A further limitation is the nature of assessment of clinical and radiographic union. The case was considered a success if clinical and radiographic union were achieved and the case was discharged with no further visits recommended. For ESF cases, this coincided with the time when frame removal occurred. These are all subjective assessments and are also strongly influenced by when the client returns the animal for re-examination. While this will not have affected our primary finding (complication rate), this subjectivity may have affected the findings regarding length of follow-up or number of revisits required.
In conclusion, this study demonstrated a significant difference in complication rates between the methods of stabilization of feline diaphyseal tibial fractures with ESF resulting in a significantly higher complication rate compared to ORIF. It also demonstrates that postoperative complications are more common when more complex ESF frames are used. Based on these results, it may be prudent to select ORIF for stabilization of feline tibial fractures wherever practical. Where it becomes necessary to use ESF, the owner should be completely informed regarding complication rates, length of follow-up required, and projected costs associated with this. Future studies are indicated to investigate the roles that ESF frame destabilization and minimally invasive plate osteosynthesis may have in reducing the complication rates associated with these challenging injuries.