Materials and Methods

Animals—The study was conducted at 2 veterinary teaching hospitals and 5 private referral hospitals located in Canada (n = 6) and the United States (1). All study procedures were approved by the Ontario Veterinary College Animal Care Committee. All dogs that were admitted to the hospitals for a TPLO between September 2012 and March 2014 were eligible for study enrollment. Dogs that underwent 2 separate TPLO procedures on different dates were considered independent cases. Dogs that underwent bilateral TPLO were classified as a single case.

Sample collection and processing—For each dog at the time of admission, each of the following areas was individually swabbed with a sterile aerobic culture swab^a: 1 naris, pharynx, rectum, and skin at the TPLO site. All swab specimens were refrigerated to ensure bacterial viability and shipped to the microbiology laboratory at the Ontario Veterinary College every 2 weeks throughout the study period. A preoperative questionnaire was administered to the owner of each dog to obtain information regarding exposure to antimicrobials prior to admission, presence of comorbidities (eg, atopic dermatitis and hypothyroidism), and amount of interaction with other dogs (eg, other dogs in the household or visits to dog parks). During the postoperative recheck examination (6 to 8 weeks after surgery), 3 of the participating hospitals obtained a second set of swab specimens as described to determine the postoperative MRSP carriage status of study dogs.

Bacteriologic culture—Each swab was placed in an individual test tube that contained an enrichment broth consisting of tryptone (10 g/L), sodium chloride (75 g/L), d-mannitol (10 g/L), and yeast extract (2.5 g/L) and incubated at 35°C for 24 hours. Then, approximately 10 μL of broth from each tube was inoculated onto mannitol salt agar that contained oxacillin (2 μg/mL) and incubated at 35°C for 48 hours. Bacterial colonies that had a yellow color as a result of metabolism of mannitol salt were suspected to be S pseudintermedius and subcultured onto Columbia blood agar that contained 5% sheep blood and incubated at 35°C for 24 hours.

Isolates were presumptively identified as S pseudintermedius on the basis of colony morphology, gram-positive appearance, positive catalase and coagulase reactions, and negative results on a Staphylococcus aureus latex agglutination test.^b Bacterial DNA was extracted^c from presumptive S pseudintermedius isolates and analyzed with an S pseudintermedius–specific PCR assay as described.²⁶ Positive and negative controls were included in each PCR assay. Resistance to methicillin was confirmed by the use of a penicillin-binding protein 2a latex agglutination test.^d The genetic sequences of isolates that were confirmed to be MRSP were further characterized by mec-associated dru typing as described,²⁷ with dru repeats and types assigned by the use of an established database.^e

Data collection and analysis—Data recorded for each dog included timing and dosage of preoperative, intraoperative, and postoperative antimicrobials; duration of surgery; duration of anesthesia; whether an SSI developed; bacteriologic culture results from the SSI (when available); and whether the implant had to be removed. An SSI was defined on the basis of criteria established by the US CDC.²⁸ Active surveillance for SSIs was performed by contacting the owner of each dog by telephone between 30 days and 1 year after the TPLO. Information from the telephone interview combined with the information obtained during the postoperative recheck examination was used to identify dogs with an SSI.

Dogs were classified as MRSP carriers if MSRP was cultured from any of the 4 sites swabbed. All outcomes of interest were dichotomous and included preoperative and postoperative MRSP colonization, development of an SSI caused by MRSP within 30 days after TPLO, and development of an SSI caused by any pathogen within 30 days after TPLO. Univariable analysis of the association between each independent variable and each outcome was performed by use of the Pearson χ² test, Fisher exact test, or logistic regression analysis. Variables with P < 0.20 on univariable analysis were selected for multivariable analysis. For each outcome of interest, stepwise backward logistic regression analysis was performed, and variables with P < 0.05 were not retained in the model unless they were deemed to be confounders. Any variable that resulted in a > 20% change in the coefficients for the remaining independent variables when it was removed from the model was considered a confounder and retained in the final model. All possible 2-way interactions were evaluated. Only 2-way interactions and variables with P < 0.05 and confounders were retained in the final model. Because of the small number of events for each outcome, multiple subset logistic regression was also conducted as described,²⁹ and the results were compared with those of the stepwise backward logistic regression. Pearson residuals were examined to identify outliers, and outlier values were further investigated to ensure that the outlier was not the result of an error made during data collection or entry. Bayesian information criterion and AIC were used to assess how well each model fit the data.

Results

Dogs—Five hundred forty-nine dogs with a mean ± SD age of 5.5 ± 2.7 years (range, 11 months to 13.1 years) and weight of 37.4 ± 11.8 kg (82.3 ± 26.0 lb; range, 5.6 to 81 kg [12.3 to 178.2 lb]) were enrolled in the study. The study population consisted of 277 (50.5%) spayed females, 250 (45.5%) castrated males, 11 (2%) sexually intact females, and 9 (1.6%) sexually intact males; the sex was not indicated in the medical records of 2 patients. Seventy-four breeds were represented, with the most common being mixed breed (n = 113 [20.6%]), Labrador Retriever (101 [18.4%]), and Golden Retriever (38 [6.9%]). The TPLO was performed on the right hind limb of 256 (46.6%) dogs, left hind limb of 258 (47.0%) dogs, and bilaterally in 26 (4.7%) dogs; information regarding the limb on which the TPLO was performed was unavailable for 9 (1.6%) dogs.

MRSP colonization—Methicillin-resistant S pseudintermedius was isolated before TPLO from 24 of 549 (4.4%) dogs. Of those dogs, MRSP was isolated from the pharynx of 12 (50%), nares of 6 (25%), rectum of 10 (42%), and skin of 5 (21%). Three dogs had MRSP isolated from 3 of the 4 sites swabbed (pharynx, nares, and rectum [n = 1]; pharynx, rectum, and skin [1]; and pharynx, nares, and skin [1]), and 4 dogs had MRSP isolated from 2 sites swabbed (pharynx and rectum [n = 2], pharynx and nares [1], and rectum and skin [1]). Swab specimens were obtained during the postoperative recheck examination from only 193 of the 549 (35.2%) study dogs and only 12 of the 24 (50%) dogs from which MRSP was isolated before TPLO. Methicillin-resistant S pseudintermedius was isolated from 17 (8.8%) of those dogs, including 10 of the dogs from which MRSP was isolated before TPLO. Of those dogs, MRSP was isolated from the pharynx of 7, nares of 2, rectum of 5, and skin of 6. Three dogs had MRSP isolated from 2 of the 4 sites swabbed (pharynx and nares [n = 2] and pharynx and rectum [1]). Dogs from which MRSP was isolated before TPLO (10/12) were significantly (P < 0.001) more likely to have MRSP isolated after TPLO than were dogs from which MRSP was not isolated before TPLO (7/181). For the 3 hospitals that obtained postoperative swab specimens, the prevalence of MRSP isolates did not differ significantly between preoperative and postoperative swab specimens.

The most common MSRP dru types were dt9a, dt10h, and dt11af. Nine of the 10 dogs that were identified as MRSP carriers both before and after TPLO harbored the same dru type, and 9 of 10 dogs from which MRSP was isolated from multiple sites harbored the same dru type at all MRSP culture–positive sites.

The preoperative and postoperative culture results were pooled to determine the site-specific sensitivity of swab specimens for identifying MRSP carriers. Collectively, MRSP was isolated from preoperative, postoperative, or both preoperative and postoperative swab specimens of 41 dogs. The specificity for identifying MRSP carriers was 44% (18/41) for swab specimens obtained from the pharynx, 19.5% (8/41) for swab specimens obtained from the nares, 36.6% (15/41) for swab specimens obtained from the rectum, and 29.3% (12/41) for swab specimens obtained from the skin.

Surgical outcome—Information was available for all 549 dogs at 30 days after TPLO and 223 dogs at 1 year after TPLO. All dogs were administered perioperative antimicrobials, so this factor was not further evaluated. Three hundred ninety-eight of 549 (72.5%) dogs were administered postoperative antimicrobials for a median of 10 days (range, 12 hours to 21 days). Thirty-five (6.4%) dogs developed an SSI within 30 days after surgery, 1 dog developed an SSI at 3 months after surgery, and 1 dog developed an SSI at 10 months after surgery. Therefore, the overall SSI incidence rate within 1 year after TPLO for the study population was 6.7% (37/549). Among the 7 participating hospitals, the SSI incidence rate within 30 days after TPLO ranged from 0% to 15.7% (mean, 5.9%). The TPLO implant was removed from 25 of the 37 (67.6%) dogs that developed an SSI.

Specimens obtained from the SSI were submitted for bacteriologic culture for 32 of the 37 (86.5%) dogs that developed an SSI after TPLO. One or more bacterial isolates were cultured from 27 (84.4%) of those specimens and included MRSP (n = 11 [40.7%]), S pseudintermedius (8 [29.6%]), Streptococcus spp (4 [14.8%]), S aureus (2 [7.4%]), MRSA (2 [7.4%]), and Enterococcus faecalis, Enterococcus faecium, Escherichia coli, Actinomyces spp, and Pasteurella canis (1 [3.7%] each).

Variables associated with preoperative and postoperative MRSP colonization and development of an SSI after TPLO—Descriptive statistics for the variables assessed by univariable analyses for association with each outcome of interest and the resulting P values were summarized (Table 1). The final multivariable models for the 4 outcomes of interest (preoperative and postoperative MRSP colonization and development of an SSI caused by MRSP or any pathogen within 30 days after TPLO) were likewise summarized (Table 2). When site-specific preoperative MRSP colonization was substituted for preoperative MRSP colonization in the multivariable model for postoperative MRSP colonization, preoperative MRSP colonization of the pharynx (OR, 58.9; 95% CI, 5.42 to 641; P = 0.001) was identified as a risk factor for postoperative MRSP carriage; however, that model had a higher AIC and did not fit the data as well as the model that included preoperative MRSP colonization. Conversely, the multivariable model for development of an SSI caused by MRSP within 30 days after TPLO that included site-specific MRSP colonization had a lower AIC and fit the data better than did the model that included preoperative MRSP colonization (OR, 14.8; 95% CI, 4 to 54.7; P < 0.001).

Individual hospitals were not included in the final multivariable model for any outcome except the development of an SSI caused by any pathogen within 30 days after TPLO. The 3 hospitals (A, E, and G) identified as risk factors for the development of an SSI within 30 days after TPLO either did not administer postoperative antimicrobials or only administered antimicrobials for < 24 hours after TPLO in most cases. Administration of antimicrobials after TPLO appeared to have a protective effect, with patients that were not administered postoperative antimicrobials being 3.5 times as likely to develop an SSI after TPLO, compared with patients that received postoperative antimicrobials. When preoperative MRSP colonization was replaced by site-specific MRSP colonization, preoperative MRSP colonization of the skin (OR, 26.5; 95% CI, 3.29 to 214; P = 0.002) was identified as a risk factor for development of an SSI within 30 days after TPLO; however, the multivariable model with preoperative MRSP colonization had a smaller AIC and fit the data better.

Evaluation of the effect of perioperative antimicrobial administration on the development of an SSI after TPLO was hampered because the protocols for antimicrobial administration in patients undergoing TPLO were fairly homogenous among the 7 participating hospitals. However, at 1 hospital, only 7 of 79 (8.9%) study dogs that were administered postoperative antimicrobials developed an SSI, whereas 18 of 74 (24.3%) study dogs that did not receive postoperative antimicrobials developed an SSI. This indicated that antimicrobial administration after TPLO was protective (OR, 0.3; 95% CI, 0.12 to 0.76; P = 0.015) at that hospital.

Discussion

Results of the present study indicated that preoperative MRSP colonization was a risk factor for the development of an SSI after TPLO in dogs. The prevalence (4.4% [24/549]) of dogs with preoperative MRSP colonization in the present study was similar to that reported by investigators of other studies.^15,16,30 In the present study, the final multivariable models for all outcomes (preoperative MRSP colonization and development of an SSI caused by MRSP or any pathogen within 30 days after TPLO) except postoperative MRSP colonization included Bulldog as a risk factor. The reason for this finding was not investigated. Bulldogs are predisposed to skin-fold dermatitis,^31,32 and MRSP is frequently isolated from the skin of dogs with active or recent pyoderma.¹⁶ However, pyoderma was not associated with any of the outcomes in the present study, and none of the Bulldogs in the study had active pyoderma at the time of surgery. Although preoperative antimicrobial administration and pyoderma were not identified as risk factors for any of the outcomes in the present study, it is possible that at least some of the Bulldogs had pyoderma that was undiagnosed.

The final multivariable models indicated that weight was also a risk factor for MRSP colonization; the risk of preoperative and postoperative MRSP colonization increased by 9% and 7%, respectively, for each 1-kg increase in body weight. The reason for this association and whether it is a function of increased lean mass or obesity are unknown.

In the present study, MRSP was most frequently isolated from swab specimens obtained from the pharynx, a finding that was similar to results of a study¹⁷ involving dogs admitted to veterinary hospitals in Germany. Isolation of MRSP from the 4 sites evaluated in the present study was variable, and although the site-specific sensitivity of culture results for identifying dogs carrying MRSP was highest for pharyngeal swab specimens, it was only 44%. Because none of the sites (pharynx, nares, rectum, and skin) had a high individual sensitivity for identifying MRSP carriers, it may be prudent to culture all 4 sites and evaluate the results collectively to maximize the likelihood of identifying MRSP carriers.

Results of other studies^16,17,30 suggest that hospitalization is a risk factor for MRSP colonization as is antimicrobial administration.¹⁷ All of the dogs of the present study were hospitalized and administered antimicrobials perioperatively; however, the prevalence of MRSP carriers after TPLO did not differ significantly from the prevalence of MRSP carriers before TPLO for any of the 3 hospitals that collected postoperative swab specimens for bacteriologic culture. Sources of MRSP acquisition for carriers, although not specifically investigated in the present study, could include exposure to environmental contamination inside or outside of the hospital and the contaminated hands or clothes of healthcare providers. Ten of the 12 dogs that were identified as MRSP carriers before TPLO and also had postoperative swab specimens collected remained culture positive for MRSP, and the majority (9/10) of those dogs were shedding the same strain of MRSP both before and after TPLO. This finding suggested that MRSP can be isolated from carrier animals for an extended period of time. In a study³³ of 12 MRSP culture–positive dogs, 2 dogs remained culture positive and 5 dogs were intermittently culture positive for 6 months, and the remaining 5 dogs became culture negative at various times during the 6-month observation period and remained so for the duration of the study. In that study,³³ MRSP was cultured from the environment of 4 households that did not contain MRSP carrier dogs. Thus, it appears that MRSP carriers may shed the bacterium for long periods, which has implications for infection control practices in veterinary hospitals, and the apparent positive association between MRSP carriage and risk of SSI is concerning.

The overall mean incidence rate of SSI following TPLO in dogs among the hospitals of the present study (6.7%; range, 0% to 15.7%) was consistent with that (range, 2.5% to 15.8%) reported by investigators of other studies.^3,11 The variation in the SSI incidence rates among the hospitals of the present study was striking, and the reason for this remains unclear. The hospital with the lowest SSI incidence rate following TPLO administered postoperative antimicrobials to all patients, whereas the hospital with the highest SSI incidence rate following TPLO only administered postoperative antimicrobials to select patients for a limited time and had low SSI incidence rates for other procedures and an active infection control program in place. Throughout the present study, SSIs developed sporadically and were not associated with any identifiable outbreaks.

In the present study, dogs from which MRSP was isolated before TPLO were at increased risk for developing an SSI, compared with dogs from which MRSP was not isolated in a manner similar to human carriers of MRSA.^18,19,23 More specifically, dogs from which MRSP was isolated from the nares and rectum before TPLO were 14 and 13 times, respectively, as likely to develop an SSI caused by MRSP within 30 days after TPLO. Surprisingly, isolation of MRSP from the pharynx was not identified as a risk factor for the development of an SSI caused by MRSP within 30 days after TPLO, even though the pharynx was the site from which MRSP was most frequently isolated. It is possible that MRSP from the pharynx is less likely to cause an SSI than is MRSP from the nares or rectum because the pharynx is more internally located than either the nares or rectum. However, isolation of MRSP from the pharynx before TPLO was positively associated with isolation of MRSP from the pharynx after TPLO, which suggested that the pharynx may be the primary site of MRSP colonization in carrier dogs.

In human medicine, preoperative screening of patients for MRSA prior to elective surgeries and the implementation of presurgical decolonization therapy for patients with MRSA-positive results have been both clinically and financially beneficial.^24,25,34,35 In 1 study,³⁴ 20 of 741 (3%) human patients who underwent joint arthroplasty without being screened or treated for MRSA before surgery developed an SSI, whereas only 17 of 1,440 (1%) similar patients who were screened and treated for MRSA as necessary before surgery developed an SSI. It is unknown whether preoperative screening and decolonization strategies for MRSP would be beneficial for dogs. An eradication program for any pathogen must be both effective and efficient to be clinically useful. In the case of MRSP, carriers should be identified quickly and the screening test used should have high sensitivity and specificity to minimize erroneous test results.³⁶ In a study³⁶ conducted to assess screening methods for MRSA in human patients, a real-time PCR assay provided the quickest results and had the highest sensitivity when compared with a broth-enriched culture method; however, although the amount of labor required for each method was approximately the same, the real-time PCR assay was 2.27 times more expensive than the broth-enriched culture. Unfortunately, a real-time PCR assay for identification of MRSP is currently unavailable. Development of a real-time PCR assay or other rapid assay for identification of MRSP would facilitate MRSP screening programs because the 48 to 72 hours required to attain bacteriologic culture results makes it inconvenient and logistically challenging to screen veterinary patients for MRSP prior to surgery, especially if owners must travel a substantial distance to bring the patients to a surgical facility for collection of the culture specimens.

Various MRSA decolonization protocols have been described and studied in human patients, but to our knowledge, comparable information is unavailable for veterinary patients. Development of decolonization protocols for MRSP in dogs is challenging because most MRSP isolates are susceptible to only a limited number of antimicrobials, it is impossible to topically treat the nasal passages or pharynx of most dogs, and many dogs have MRSP colonization at multiple sites. There is also concern that decolonization protocols could facilitate the development of multi-antimicrobial–resistant bacterial isolates. An alternative to decolonization protocols for MRSP carriers would be to modify the perioperative antimicrobials administered. For example, the addition of amikacin to the preoperative antimicrobial regimen for patients with amikacin-susceptible MRSP might be practical and minimize the pressure for development of antimicrobial resistance. However, objective assessment of such an alternative protocol is necessary before it is adopted for routine perioperative patient management.

Results of the present study and other studies^2,5,6,14 indicated that dogs administered antimicrobials after TPLO were less likely to develop an SSI than were dogs that were not administered postoperative antimicrobials. In human medicine, current guidelines recommend that no or only short-term (≤ 24 hours) antimicrobials be administered to patients following surgical procedures similar to TPLO,³⁷ although that practice may contribute to antimicrobial resistance, additional morbidity, and increased treatment costs.^38–40 Human patients differ from veterinary patients, especially in terms of pathogen exposure and extent of patient care. A thorough assessment of the effects of various postoperative antimicrobial protocols on the development of SSIs after TPLO in dogs is warranted.

A limitation of the present study was the amount of time that lapsed between acquisition of specimens and bacteriologic culture for MRSP.⁴¹ Specimens were shipped to the laboratory every 2 weeks. Thus, specimens were stored for variable times before being processed for bacteriologic culture, which may have negatively affected bacterial viability and resulted in underestimation of the prevalence of MRSP carriers.

Findings of the present study indicated that dogs colonized by MRSP prior to TPLO were at increased risk of developing an SSI, compared with dogs that were not MRSP carriers. The Bulldog breed also appeared to be a risk factor for preoperative MRSP colonization and development of an SSI within 30 days after TPLO, whereas postoperative antimicrobial administration appeared to be protective against the development of SSIs. Methicillin-resistant S pseudintermedius was most frequently isolated from the pharynx, nares, and rectum of the study dogs, and specimens for bacteriologic culture should be obtained from multiple body sites when screening dogs for MRSP. Preoperative screening of dogs for MRSP and the development and implementation of MRSP decolonization protocols for MRSP carriers prior to surgery would likely be beneficial, but further investigation is warranted.