Perioperative risk factors for surgical site infection in tibial tuberosity advancement: 224 stifles

Summary Objective: To examine perioperative factors affecting surgical site infection (SSI) rate following tibial tuberosity advancement (TTA). Study design: Retrospective case series. Sample population: 224 stifles in 186 dogs. Methods: Medical records of dogs that underwent TTA in a single institution were reviewed. Information on signalment anaesthetic and surgical parameters, as well as occurrence of SSI was recorded. Dogs were followed for a minimum of three months postoperatively. The association between perioperative factors and SSI was assessed using Chi-squared tests and binary logistic regression. Results: The prevalence of SSI was 5.3% (12/224 TTA). Surgical time (p = 0.02) and anaesthesia time (p = 0.03) were significantly associated with SSI. For every minute increase in surgical time and anaesthesia time, the likelihood of developing SSI increased by seven percent and four percent respectively. The use of postoperative antimicrobial therapy was not significantly associated with lower SSI (p = 0.719). Implants were removed in 1.3% of cases (3/224 TTA). Conclusions: The findings of this study suggest that increased surgical and anaesthesia times are significant risk factors for SSI in TTA, and that there is no evidence that postoperative prophylactic antimicrobial therapy is associated with SSI rate.

: nachoC@htzpatr~ckreferrals.co.uk cruciate ligament rupture is a sigcause of morbidity in dogs, leading atrophy and osteoar-1,2). The underlying aetiologies of cruciate ligament rupture are and a myriad of treatons have been proposed (3)(4)(5)(6)(7)(8). many surgical procedures deagement of cranial crue ligament rupture, aiming at improvthe stability of the stifle. In the last two cades, the attention has shifted from re-placement of the cranial cruciate ligament to providing stability through altering the stifle biomechanics by tibial osteotomy. One of the more recently developed surgical techniques, the tibial tuberosity advancement (TTA), has been shown to reduce tibial translation in the cranial cruciate ligament-deficient stifle and provide good clinical outcomes (9).
The prevalence of surgical site infection (SSI) in clean orthopaedic surgical procedures in dogs and cats is 2.5 -4.8% (10)(11)(12)(13). Many risk factors for SSI have been identified including prolonged surgi-cal time and pre-anaesthetic clipping (1 1). Previous studies have reported a SSI rate of 2.6 -7.2% for TTA surgery (9,14,15). A study involving tibial plateau levelling osteotomy and extracapsular lateral suture showed that tibial plateau levelling osteotomy had a significantly higher prevalence of SSI (8.4%) when compared to extracapsular lateral suture (4.2%) (p = 0.01) (16). The study also found that skin sutures and postoperative prophylactic antimicrobial therapy were associated with a lower risk of SSI (16). Other studies of tibial plateau levelling osteotomy found that an increased body weight and being male were associated with increased risk of SSI whereas the Labrador Retriever breed, the use of postoperative prophylactic antimicrobial therapy and the application of a locking bone plate were associated with a lower risk of SSI (1 7-19).
In this study, we examined the perioperative and postoperative factors of the TTA procedure that might be associated with SSI in a university teaching hospital. We also examined the association between the use of postoperative prophylactic antimicrobial therapy and SSI in TTA. Our null hypothesis was that the administration of postoperative prophylactic antimicrobial therapy was significantly associated with lower SSI in dogs undergoing the TTA procedure. -Purulent discharge from deep ~nc~sion -An abscess or other evidence of ~nfection involving the deep incision is found on direct examination, during revision surgery or by histopathology and radiologic examinat~on -A deep incision spontaneously dehisces or is deliberately opened by a surgeon in cases w~th pyrexla, localized pain or tenderness, unless ult of the ~ncision is n e g a t i v ecases provided that the selection criteria were met. In order to eliminate variables that may have led to an increased risk of SSI, cases that had antimicrobial therapy administration within seven days prior to surgery for other infectious conditions were excluded from the study. Cases that might have infectious processes that were not treated with antimicrobial therapy were not excluded. Cases that had less than three months postoperative follow-up or had traumatic events that could contaminate the surgical site, such as open tibial tuberosity fracture resulting in an opened surgical wound, wereeicluded from this study, Information collected from the medical records included age and breed of dogs, affected stifle joint (left or right limb), surgical time, anaesthesia time, anaesthesia induction agent, type of pre-and intra-operative prophylactic antimicrobial therapy used, the interval between first prophylactic antimicrobial therapy dose given and the first slun incision, perioperative nadir body temperature, whether postoperative prophylactic antimicrobial therapy was used, the types of postoperative prophylactic antimicrobial therapy if prescribed, signs of SSI, bacterial culture and sensitivity result from SSI, and the interval between surgery and the development of SSI.

Data collection
Vet Comp Orthop Traumatol 31201 5 The surgical time was defined as the interval from the first skin incision to the final skin closure. The anaesthesia time was defined as the time from induction to the extubation. The perioperative nadir temperature was defined as the lowest recorded body temperature under anaesthesia and during the immediate recovery period.
Stifles that developed SSI were further categorized into superficial incisional SSI and deep incisional SSI as defined by the Centers for Disease Control and Prevention (FTable 1) (20). All of the samples collected for bacteriology were incubated for at least 48 hours on sheep blood and MacConkey agar prior to bacterial identification and antibiotic susceptibility testing. If an anaerobic bacterium was suspected, the sample was also incubated on horse blood agar.
All the immediate postoperative radiographs were assessed by one of the authors (IC) for any technical errors of the surgical procedures as a way to assess the association between implant instability and SSI. The criteria for assessment of technical errors was adopted from a previous study that assessed the risk factors associated with tibial tuberosity fractures in TTA; these criteria were osteotomy shape, cage position and plate position (21). Cases that had any one or more of the technical errors of undesirable osteotomy shape and cage and plate position were considered to be positive for technical error. I Pre-, peri-and postoperative r a All dogs were premedicated with ative analgesia included the adrnini of methadone (0.3 -0.5 mglkg required for up to 24 hours, which was lowed by buprenorphinen (10 -20 pgl ,6 -12 hours. Meloxicam0 (0.1 mg/kg per os), firocoxibp (5 mglkg sid per 0s) robenacoxibq (1 -2 mglkg sid per os), s prescribed for seven to 21 days postopively. Postoperative prophylactic anticrobial therapy with cephalexiner (15lmglkg bid per 0s) or amoxicillin with anic acid"l0 -20 mglkg bid per 0s) s prescribed in some cases for five to 10 s, depending on the preference of the ary surgeon. In general, dogs were talized for one to two days and were examined at six to eight weeks postopervely or when complications occurred.
hs of the affected ion or general anaesthesia prior to ry as part of the surgical planning. e sedation and anaesthesia protocol for the stifle radiographs were at the n were closed in layers; the use of skin ures or skin staples was at the discretion f the primary surgeon.
gistic regression was used to examine asciations between clinical data and SSI.
ariable screening of all variables was ucted, including dog identification, as random effect to account for the clusterof surgical procedures within dogs. All iables with p-values t0.25 from the in-   No 69 *initial deep tissue swab; +bacteriology from implant after 2 weeks of amoxicillin-clavulanic acid; *after initial course of antimicrobial therapy; MDR = multi-drug resistant, 'resistant to three or more antimicrobial classes (53); XDR = extensively drug-resistant, bacterium remains susceptible to only one or two antimicrobial classes (53).
itial univariable screen were considered to be eligible for inclusion in a multi-level multivariable logistic regression model. A forward selection process was used during the building of the multivariable model and variables were retained within the model if the p-value associated with that variable remained t0.05. T-tests or Mann-Whitney tests (dependent on evidence of a lack of normality in the data) were used to examine associations between individual clinical variables.

Result
The medical records of 192 dogs with 230 TTA procedures were reviewed. Two hundred and twenty-four stifles from 186 dogs met the inclusion criteria. Six dogs were excluded from the study due to antibiotic administration for a condition unrelated to the TTA within seven days prior to the surgery (n = 3), open fracture of the tibid tuberosity (n = I), self-trauma from patient which resulted in removal of all skin sutures (n = I), and death due to septic peritonitis secondary to gastric ulceration 11 days postoperatively (n=l). One hundred and seventeen TTA were performed in male dogs (52%) and 107 TTA were performed in female dogs (48% gical time in dogs that developed an SSI utes). Twelve dogs had SSI giving a prevalence was 130 minutes (+ 39 minutes), which was F 5.3%. The breeds of dogs that developed significantly longer than the dogs that did SSI were: Boxer (n = 4), Labrador Renot develop an SSI (102 minutes + 25 mingroup developed SSI, and three out of iever (n = 2), Springer Spaniel (n = l), utes) (p = 0.02). olden Retriever (n = l), Rottweiler (n = There were three superficial incisional I, Douge De Bordeaux (n = I), and Border SSI and nine deep incisional SSI. Of the ollie (n = 1). The Springer Spaniel dog nine stifles with deep incisional SSI, four ad TTA performed bilaterally with an instifles had draining sinuses, three had lrval of 183 days between surgical proswelling and pain at the surgical sites, two :dures and developed SSI bilaterally. Surhad septic arthritis, one stifle had excessive .cal site infections were found in nine radiographic signs of periosteal proliferght stifles and three left stifles. The mediation of the surgical site, and one stifle had I time from TTA to the development of signs of radiolucency around the TTA fork.   ment and two weeks of amoxicillin-clavulanic acid and enrofloxacin. The third implant removal was not associated with SSI; the TTA cage with a broken cage screw was removed due to impingement of the patellar ligament.

Regression analysis
Univariable analysis showed that surgical time was significantly longer in dogs that developed SSI than those that did not (p = 0.02; odds ratio (OR): 1.07; 95% confidence interval (95%CI): 1.01 -1.14) as was anaesthesia time (p = 0.03; OR: 1.04; 95%CI: 1.004 -1.08) (, Table 4). For every minute increase in surgical time and anaesthesia time, the likelihood of developing SSI increased by seven and four percent, respectively. When comparing two dogs at the 75th (120 minutes) and 2 5~ (86 minutes) percentiles for surgical time, a dog with a surgical time of 120 minutes was 10 times (1.07(l~O-~~)) more likely to develop SSI compared to a dog with a surgical time of 86 minutes.
It was not possible to build a model that included more than one variable. Surgical time and anaesthesia time were highly correlated (Spearman correlation coefficient = 0.8) (,Appendix Figure 1: available online at www.vcot-online.com) and none of the other variables were statistically significant when either surgical time or anaesthesia time were included as starting points in the multivariable model building process.

Discussion
In this study, we found that surgical time was a significant factor in the development of SSI in this cohort of dogs undergoing TTA surgery. Furthermore we have shown that the use of postoperative prophylactic antimicrobial therapy does ,sot appear to influence the incidence of SSI. As a result, we rejected the null hypothesis.

F. W. Yap et al.: Perioperative risk factors of surgical site infection in TTA
The prevalence of SSI in this study was similar to previous reports of post TTA SSI (2.6 -7.2%) (9,14,15). Despite this similarity, interpretation of these results should be performed with caution as the definition and criteria of SSI used in all the other studies were not standardized. In this study, we have used the widely adopted definition outlined by the Centers of Disease Control and Prevention (20). Adaptation of a standardized SSI definition would enable direct comparison of the results between various studies and eliminate this as a limitation.
Staphylococcus species, including multidrug resistant strains, are the most commonly isolated bacteria in reported cases of TTA and tibial plateau levelling osteotomy SSI (9,14,17,22,23). In dogs and cats, the most common isolates associated with implant-related SSI are Staphylococcus species and Streptococcus species; Staphylococcus species account for 50 -60% of infections (24). In our study, the most commonly cultured bacteria were Staphylococcus species, followed by Streptococcus species.
The prevalence of implant removal in our study was 1.3%. To the authors' knowledge, this is the first study reporting the prevalence of implant removal after TTA. In comparison, the reported prevalence of implant removal for tibial plateau levelling osteotomy ranges from 1.3 -8.5% (17, 22,23,25). Two of the stifles had implant removal due to implant associated SSI. The implants from one of these stifles cultured positive for ~lebsiella'pneumonia but it was not possible to determine whether the infection was intrdduced during or after surgery. Klebsiella pneumonia is not a commonly reported pathogen in implant associated SSI.
Surgical site infection is a commonly reported complication in the tibial osteotorny techniques for the treatment of canine cruciate disease (9,14,17,26). Administration of perioperative prophylactic antimicrobial therapy is a common practice for the tibial osteotomy techniques; the types and durations of prophylactic antimicrobial tlierapy vary greatly between practices and surgeons (18,27). Two large retrospective studies have shown that postoperative prophylactic antimicrobial therapy after surgical procedures for cranial cruciate ligament rupture significantly reduced the incidence of SSI (16,17). One of these involved only tibial plateau levelling osteotomy cases while the other involved tibial plateau levelling osteotomy and extracapsular lateral suture equally (16,17). The significant relationship between postoperative prophylactic antimicrobial therapy and lower SSI rate in these previous two studies may be attributed to the prevention of infection or to the reduction in the identification of SSI in the postoperative period due to postoperative antimicrobial therapy (16,17,28). In contrast, our study on TTA did not show a significant relationship between postoperative prophylactic antimicrobial therapy and SSI. Postoperative prophylactic antimicrobial therapy may have protective effects on factors that are applicable to tibial plateau levelling osteotomy but not to TTA, such as the potentially more extensive soft tissue dissection around the proximal tibia, the bulky profile of tibial plateau levelling osteotomy plates and the stainless steel property of the tibial plateau levelling osteotomy plates. These factors may contribute to dead space and may increase the risk of soft tissue complications (29). Titanium implants have been shown to be more biocompatible than stainless steel implants with tissues, and titanium is associated with lower implant associated infection in some studies, but not the others (30)(31)(32). As SSI can be augmented by contact necrosis of soft tissue and bone from periosteal compression, the differing degree of area of periosteal compression resulted from the application of bone plate between TTA and tibial plateau levelling osteotomy may also influence the prevalence of SSI and the protective effect of postoperative prophylactic antimicrobial therapy (33). The bion~echanical difference between TTA and tibial plateau levelling osteotovily may also explain why postoperative prophylactic antimicrobial therapy may have a protective effect against SSI in tibial plateau levelling osteotomy. The nonloclzing plate system has been associated with significantly higher SSI in dogs >50 kg; micro-motion and instability at the osteotomy site were suggested to be an important factor in this difference (18). In contrast, micro-motion across the osteotomy may not be as important a factor in m TTA as it is in tibial plateau levelling osteotomy. The osteotomy is largely not in contact with implants or adjacent bone. However, further study is warranted to sess the micro-motion at the osteotomy site rand bone-implant interface in TTA.
The majority of the published literature in human medicine show that prophylactic antimicrobial therapy after wound closure in a clean surgery is unnecessary, as additional doses have not shown benefit (34)(35)(36)(37). A large prospective study in cats and dogs (1010 surgical wounds from multiple types of surgery) also showed the lack of association of postoperative prophylactic antimicrobial therapy and SSI (p = 0.266) (1 1). Concerns with prolonged use of postoperative prophylactic antimicrobial therapy are its limited benefit in the prevention of SSI and its association with emergence of resistant bacterial strains (34,38). The current recommendation of prophylactic antimicrobial therapy in human orthopaedic surgery is that the first dose of prophylactic antimicrobial therapy begins within 60 minutes before surgical incision and be discontinued within 24 hours of the end of surgery (39).
There was no significant association between timing of first prophylactic antimicrobial therapy dose and SSI shown in our study. This is likely to be a type I1 statistical error as the number of dogs (n = 6) that received the prophylactic antimicrobial therapy after the first surgical incision was small. It is advisable that an appropriate prophylactic antimicrobial therapy effective against the likely encountered organism should be initiated within 60 minutes of the first surgical incision to achieve serum and tissue levels that exceed the minimum inhibitory concentration (39). Interestingly, some studies have found that the administration of intravenous prophylactic antimicrobial therapy four to 30 minutes prior to the time of incision resulted in lower SSI, in contrast to another study showing that the lower SSI was associated with administration of intravenous prophylactic antimicrobial therapy between 30 -60 minutes prior to skin incision (40)(41)(42). Re-administration after two half-lives of the medication is recommended intra-operatively to ensure adequate concentrations of antimicrobial in the serum and tissue ing osteotomy (18). The other half of SI ranged from 15 days to 235 days authors speculated that the early SSI as the type of procedures involved. s study showed a significant associbetween long surgical and anaestimes with SSI in TTA. The finding ificant association between long surh increased environmental contarni-E2ion of the surgical wound (49). In addiry is often associated lexity of the surgery or limited experience of the primary on. Limited surgeon experience may se the risk of surgical mistakes, rectly affect the mechanical implants and the risk of ications, such as SSI. has been associated ssociation between peritemperature and SSI is not study despite published eviedicine showing a sigassociation between perioperative rmia and higher SSI rate, delayed therapy, the lack of prospective allocation means that other factors that were not recorded may have affected the decision to prescribe postoperative prophylactic antimicrobial therapy, potentially affecting the outcome. Another limitation of this study is the relatively small number of subjects included. However we have reported very similar infection prevalence on both groups with and without prophylactic antimicrobial therapy, which suggests our result retains clinical relevance with the numbers we have reported despite the potential for a type I1 statistical error. The minimum follow-up period was three months; this might restrict the documentation of some deep ihcisional SSI cases that occur after three months. According to the Centers for Disease Control and Prevention, deep incisional infection can occur within one year after the procedure if an implant is left in place (20). The lack of speciation of some of the cultured bacteria, especially the Staphylococcal species, is also another major limitation. Methicillin-resistant Staphylococcus pseudintermedius has been recognized as a significant pathogen in companion animals recently (51). Finally, the absence of an active surveillance for SSI at the hospital may underestimate the prevalence of SSI in the TTA performed (52).
In conclusion, we have shown that postoperative prophylactic antimicrobial therapy does not appear to be a protective factor in TTA and that surgical and anaesthesia times were significant risk factors for SSI in TTA.