Susceptibility of Incisional Site Staphylococcus spp. to Three Common Perioperative Antimicrobials in Dogs Undergoing Clean Orthopaedic Procedures

Abstract Objectives The aim of this study was to evaluate the susceptibility pattern of the incisional site Staphylococcus spp. to three commonly used perioperative antimicrobials in dogs. Materials and Methods A cohort of client-owned dogs was included in this prospective observational study. After induction of general anaesthesia and hair clipping, culture samples were collected from the incisional site of healthy dogs before orthopaedic surgery. The isolated colonies were identified. The susceptibility pattern of staphylococcal isolates was determined using Clinical and Laboratory Standard Institute standards. The staphylococcal susceptibilities to amoxicillin/clavulanic acid and cefpodoxime were compared with cefazolin using two-sided equality of proportions test. Results A total of 391 bacterial isolates were identified in 115 dogs. Bacillus spp. (n = 200) and Staphylococcus spp. (n = 95) were the most common bacteria isolated. Coagulase- positive and -negative staphylococci were isolated in 44 and 38% of dogs respectively. Cefazolin, amoxicillin/clavulanic acid and cefpodoxime staphylococcal susceptibilities were 68, 77 and 76% respectively. There were no significant differences in Staphylococcus spp. susceptibility between amoxicillin/clavulanic acid or cefpodoxime and cefazolin (p = 0.19, p = 0.25). Thirty-six per cent of coagulase-negative staphylococci and 6% of coagulase-positive staphylococci isolated were methicillin resistant. Conclusion Cefazolin, amoxicillin/clavulanic acid and cefpodoxime had similar susceptibilities for staphylococcal isolates. Routine use of amoxicillin/clavulanic acid or cefpodoxime in replace of cefazolin should be practiced with caution.


Introduction
Cefazolin (CFZ) is the most commonly used perioperative antibiotic medication in human and veterinary orthopaedic surgery. It has low toxicity, high bone and joint tissue concentrations, and its use is associated with decreased rates of surgical site infections (SSI) compared with dogs that receive no perioperative antimicrobials. [1][2][3] Second-and third-generation cephalosporins or potentiated penicillin are also used as perioperative antimicrobials. 4,5 They may have the advantage over CFZ because they tend to be effective against β-lactamase producing Staphylococcus, 6 whose prevalence is increasing. 7,8 However, information about their effectiveness as perioperative prophylactic antibiotic medications for orthopaedic procedures in dogs is lacking.
Susceptibility patterns of potential bacteria at the surgical site are one of the considerations for selecting perioperative antimicrobials. 9 To the author's knowledge, studies conducted for commensal staphylococci susceptibility patterns have specifically investigated methicillin-resistant staphylococci colonization, characterized by the presence of mec A gene or oxacillin resistance interpreted by the culture and sensitivity report. 7,8 The susceptibility pattern of incisional site skin flora in dogs has not been evaluated in the literature. The purpose of the study was to evaluate the susceptibility of cutaneous staphylococci to commonly used prophylactic antimicrobials isolated from planned incision sites in dogs undergoing clean orthopaedic procedures. The null hypothesis tested was that staphylococcal isolates retrieved from planned incision sites would have no difference in susceptibility to cefpodoxime (CPD) and amoxicillin/clavulanic acid (AMC) than to CFZ.

Materials and Methods
This was a prospective observational study of healthy clientowned dogs. Dogs over a year of age that underwent clean orthopaedic surgery between November 2015 and May 2016 were eligible for inclusion. Dogs were excluded if they had received systemic or topical treatments for dermatological condition, systemic antimicrobial treatments within 4 weeks of presentation, had undergone a surgical procedure within 4 weeks of presentation, had a confirmed diagnosis of endocrine diseases or had skin lesions discovered after clipping. Signed informed owner consent was obtained. The study protocol was approved by the institutional animal care and use committee (protocol no 18309). Once enrolled, data collected included breed, sex, age, weight, urgency of surgery, classified as elective (surgeries for naturally occurring orthopaedic disease) or urgent (surgeries for fractures or luxations following a traumatic event), and anatomic site.
Dogs were anaesthetized, and the planned surgical incision site was clipped using a sterile no 40 clipper blade (Oster, Baca Raton, Florida, United States). The person preparing the site and taking the sample wore sterile gloves. Excess hair was vacuumed from the site without skin contact. A culture sample of the planned incision site was collected using a sterile contact agar plate (RODAC plate, ThermoFisher Scientific, Waltham, Massachusetts, United States) containing 5% sheep blood. 10 The entire surface of the agar plate was applied to the skin for 5 seconds. The plate was incubated in 5% CO 2 at 35°C for 24 hours, and the bacterial organisms were subcultured to obtain pure bacterial isolates. The pure bacterial isolates were used for identification and determination of antimicrobial susceptibility. Isolate identification utilized a combination of colony morphology characteristics, haemolytic reaction, Gram stain results, spot catalase testing and coagulase testing. Further characterization of isolate identity was based on results from conventional biochemical testing (API, BioMérieux, Durham, North Carolina, United States) (tubed media, catalase reaction, identification strips) or matrix-assisted time-of-flight mass spectrometry (Biotyper 3.0, Bruker, Fremont, California, United States).  11 Staphylococcus intermedius and Staphylococcus pseudintermedius identified with matrix-assisted timeof-flight mass spectrometry were reported together as Staphylococcus intermedius group (SIG). No CPD breakpoint has been established for coagulase-negative (CoNS), so the breakpoint for S. pseudintermedius was used to help determine susceptibility. Oxacillin-resistant isolates were reported to be resistant to methicillin and all other β-lactam antimicrobials, regardless of individual β-lactam antimicrobial susceptibility testing result. 11 Isolates interpreted to have intermediate susceptibility were considered resistant for statistical analysis. Descriptive statistics were calculated for all measured variables and reported as number (per cent), median (interquartile range) or mean (standard deviation); a combined test of skewness and kurtosis was used to evaluate normal distribution of continuous variables. Population rates of staphylococci susceptible to CFZ, CPD and AMC were estimated with surrounding exact binomial 95% confidence intervals (CI). Two-sided equality of proportions test was used to test the null hypotheses that the rate of staphylococci susceptibility to CFZ did not differ significantly from the rate of susceptibility to CPD or from the rate of susceptibility to AMC. Antimicrobial susceptibilities of CFZ, CPD and AMC were compared across staphylococcal species using Fisher's exact tests. Logistic regression was performed to explore dog variables (age, body weight, sex, anatomic surgical site and urgency of surgery [classified as elective or emergent] ) associated with isolation of CFZ-resistant staphylococci. All tests were two-sided, with p < 0.05 considered significant. Analyses were performed with computer software (Stata Statistical Software: Release 14, College Station, Texas, United States).
Staphylococcus spp. were isolated from culture samples of 83/115 (72%) dogs. Twelve dogs grew two different staphylococcal isolates, and the remaining 71 dogs grew a single staphylococcal isolate. Nine isolates identified as S. intermedius and 38 S. pseudintermedius isolates were reported as SIG. The prevalence of bacterial isolates is shown in ►Table 1.
On univariable logistic regression analysis, none of the following factors were significantly associated with growing a CFZ-resistant Staphylococcus spp.: age, sex, body weight, urgency of surgery or anatomic surgical site.
Among the 95 staphylococcal isolates, 65 (68%, 95% CI: 59.1-77.8) were susceptible to CFZ. In the 12 dogs that grew two different staphylococci, both isolates were susceptible in five dogs, both isolates were resistant in two dogs, and in the remaining five dogs, one isolate was susceptible, and the other was resistant. Among the 83 dogs from which Staphylococcus spp. were isolated, 28 (

Discussion
When evaluating all bacterial isolates, Staphylococcus spp. were the second most commonly isolated bacteria from planned incision sites. The isolated staphylococci demonstrated similar susceptibility patterns to CFZ, AMC and CFD. Therefore, the null hypothesis was accepted. In vitro testing showed that dermal staphylococci isolates retrieved from planned incision sites had no difference in susceptibility to CPD and AMC than to CFZ. Breed, surgical site, age, sex, body weight, urgency of surgery or anatomic site did not have a significant effect on the CFZ susceptibility of staphylococcal isolates. In a previous study, bulldogs were identified to have a significantly higher risk of having methicillin-resistant Staphylococcus pseudintermedius (MRSP) colonization and SSI. The differences in our results when compared with this previous study are likely due to different sample sizes and sampling methods. 12 In the present study, only cases referred to one institution were included, and samples from the incision site were collected. Additionally, only dogs that received prior antimicrobial treatment within 4 weeks or with skin lesions were excluded. A more detailed screening process including information such as antimicrobial treatment status of other animals and people in the same household, and husbandry environment, may have helped capture a more accurate image on the relationship between MRSP colonization status and incisional site bacteria.
Bacillus spp. were the most common incisional site bacteria isolated. A high prevalence of Bacillus spp. at these sites has been reported in previous studies using culture-based 13 and molecular biology methods using 16s ribosomal ribonucleic acid amplification. 14 Bacillus organisms are widely distributed in the environment and found in decaying organic matter, dust, vegetables and water. Bacillus spp. rarely causes orthopaedic SSIs. Except for Bacillus cereus and Bacillus anthracis, Bacillus spp. have low virulence and SSIs caused by Bacillus spp. were only reported in immunocompromised people, 15 and SSIs caused by Bacillus spp. have not been reported in the veterinary literature.
The prevalence of coagulase-positive staphylococci (CoPS) and CoNS in this study was similar to those reported in previous veterinary studies. One study collected samples from multiple anatomic sites reported the isolation of CoPS in 37 of 50 (74%) dogs with healthy skin. 16 Other veterinary studies reported 33 to 100% isolation rates of CoNS in healthy dogs. 13,16,17 In a human study, CoNS comprised 59.6% of total isolates from planned incision sites. 18 Differences between published studies are likely due to the variation in the sampling method, and the anatomic sites sampled. One previous study showed that Staphylococcus spp. demonstrated anatomic niche preferences, and S. pseudintermedius is more often isolated from the perineum area, larynx and rectum. 12,17 Samples from these other anatomic sites may have resulted in a higher isolation rate in the present study. However, the current study aimed to investigate the susceptibility of perioperative antimicrobial against bacteria present on incision sites; therefore, no samples were taken from other anatomic sites.
Six per cent of SIG isolated were oxacillin resistant. Due to the study design, a direct relationship between the MRSP colonization status and the bacteria present on the incision site could not be determined. In one previous study, the specificity of the incisional site sample for identifying MRSP carrier status was only 29.3% 12 ; therefore, the actual number of dogs carrying methicillin-resistant SIG is likely to be higher. In addition, 36% of CoNS isolated in the present study were oxacillin resistant. This finding is similar to that in humans, where 28.6% of CoNS isolates were oxacillin resistant. 18 Although traditionally regarded to be of low virulence, recognition of the role of CoNS as pathogens in veterinary dermatology is emerging, 19 and the importance of CoNS in orthopaedic surgery should not be overlooked. Coagulase-negative staphylococci have been reported to be the causative bacteria in 12.7 to 14.2% of canine tibial plateau levelling osteotomy infections, 20,21 in 22% of human fracture fixation infections, 22 and up to 50% of human total hip arthroplasty infections. 23,24 The high oxacillin resistance rate of CoNS observed in this study may provide one possible explanation for the high proportion of CoNS infections in human and veterinary orthopaedic surgery.
Staphylococcal isolates tested in the study showed similar susceptibilities to CPD, CFZ and AMC. The result is likely due to the high proportion of methicillin-resistant staphylococci. The addition of clavulanic acid to amoxicillin and the development of CPD helped increase the inhibition of β-lactamase and therefore have an improved coverage of methicillin susceptible staphylococci. 22,25,26 However, several bacterial adaptation to β-lactamase have been discovered over the past decades, 6 with the mec A gene coded altered penicillinbinding protein 2a being the most important one. The presence of penicillin-binding protein 2a in staphylococci is responsible for the methicillin resistance phenotype and confers intrinsic resistance to all β-lactams. 27 The prevalence of multidrug-resistant staphylococcal colonization and infection is increasing, 28 but the susceptibility of incisional site Staphylococcus in dogs has not been reported to the authors' knowledge. The selection of antimicrobial for perioperative prophylactic administration should consider multiple factors, and the results of the current study should only be a reference. Degree of contamination, prior antimicrobial usage, targeted pathogenic bacteria, tissue concentration, institutional antimicrobial resistance pattern, cost and side effects should all taken into consideration. Pharmacokinetic studies to determine the optimal dosage and interval for prophylactic administration in dogs are lacking for AMC and CPD, but are available for CFZ. 1,25 However, a direct comparison of the use of different medications in dogs is lacking. Human studies showed no difference in the SSI rate between prophylactic AMC, CPD and CFZ in human hip arthroplasty, knee arthroplasty and laparoscopic gynaecology surgery. 29,30 Using the standard for interpreting cumulative antibiogram, 31 the three antimicrobials evaluated would be classified as fair to good for in vitro efficacy against incisional site staphylococci. Although the antimicrobial susceptibility may seem low, the antimicrobials evaluated in the present study still have a high susceptibility against methicillin-susceptible staphylococci. The findings of this study may help to explain a previous study that shown the positive carrier status of MRSP being a risk factor for the development of SSI after tibial plateau levelling osteotomy. 12 Best practices to prevent SSI in patients colonized with a resistant staphylococci remain unknown. Preoperative decontamination for human patients with methicillin-resistant Staphylococcus aureus colonization has been shown to be effective in preventing SSIs. 32 Comparable information is unavailable in veterinary medicine and would require further investigation.
This study had several limitations. The confidence intervals around susceptibility rates and rate differences suggest that significant differences could yet exist between antimicrobials, although such modest differences would need to be correlated with SSI rates to determine their clinical significance. In this study, short-and long-term patient follow-up was not conducted, and the relationship between skin culture status and SSI was not examined. Multi-site research over a broader geographic area might provide additional information to guide clinical practice. Dogs in this study had no pathologic skin lesions around the incision site, and cultures were collected before the skin was surgically scrubbed. Bacteria seen on the culture may have included bacteria residing on the residual hair after clipping. Additionally, the isolation rate and susceptibility to CFZ, AMC and CFD after skin closure may have been different since perioperative antimicrobial prophylaxis is used in an effort to decrease total bacterial load throughout the procedure. Only conventional agar-based culture methods were used to detect bacteria. Recent advancements in molecular biology allowed bacterial identification using gene sequencing, 33 which provides more detailed qualitative information to identify bacterial diversity. The currently used culture-based methods, although limited in breadth for the types of organisms detected and host-microbe interaction, have the advantage of quantifying absolute cell abundances of culturable microbes.
In conclusion, we found that 68% of staphylococci isolated from planned, clean, orthopaedic incision sites to be suscep-tible to CFZ. Cefpodoxime and AMC did not provide additional susceptibility coverage for isolated staphylococci. Based on the result of the present study, the lack of pharmacokinetic data, unknown added benefit for preventing SSI and antimicrobial stewardship, routine use of AMC or CPD in replace of CFZ should be practiced with caution.