Introduction

Staphylococcus pseudintermedius is the bacterial pathogen most often isolated from lesions associated with superficial bacterial folliculitis in dogs. A common clinical concern with S pseudintermedius is the potential for methicillin and multidrug antimicrobial resistance.¹ In a clinical setting, if empiric systemic antimicrobial therapy fails to improve clinical symptoms, a bacterial culture and susceptibility is indicated.²

A range of systems are available for antimicrobial susceptibility testing, and the system used is laboratory dependent. The potential methodologies include broth microdilution (BMD), disc diffusion (DD), and commercial systems such as Sensititre, eTests, Vitek, and BIOMIC.³ The system used for antimicrobial susceptibility testing can influence the results even if performed within the same laboratory.^4,5 Some of these commercial systems have been shown to have variable results for β-lactam susceptibility when evaluating Streptococcus pneumoniae.⁴ When evaluating S pseudintermedius with BMD and DD, there was a high specificity (97.1–100%) with variable sensitivity (35.7–98.9%). Poor sensitivity was observed when evaluating amoxicillin-clavulanic acid (51.5%), cephalothin (43.6%), and cefoxitin (35.7%) with DD testing.⁶ Comparison of BMD, DD, and Vitek-2 with canine S pseudintermedius in Italy within one laboratory showed reliable performance of the DD and Vitek-2 compared with BMD. One problem identified in that study was the minimum inhibitory concentration (MIC) ranges for doxycycline and amikacin do not match the Clinical and Laboratory Standards Institute (CLSI) clinical breakpoints for canine S pseudintermedius.⁵

Another aspect of antimicrobial susceptibility testing is laboratory variability. An older study revealed a threefold dilution range in the MIC that could alter the result from sensitive to resistant and vice versa. This prior investigation showed that inoculum size, incubation time, temperature, and other environmental factors affect variability.⁷ Additionally, another study using a single strain of Staphylococcus aureus tested at 10 different laboratories revealed a 55% variability in interpretation of the MIC. Therefore, standardization of the laboratory practices was recommended to decrease this variability.⁸ A more recent analysis from 2018 with S aureus showed much lower variability between laboratories.⁹ Similar evaluations with S pseudintermedius have not been undertaken.

The objective of this study was to evaluate the culture and susceptibility results of a homogeneous sample submitted to four different veterinary laboratories, especially with more standardized practices instituted among the larger veterinary laboratories. We hypothesized the bacterial culture and susceptibility results within the evaluated facilities would correlate with high agreement for antimicrobial susceptibility.

Materials and Methods

Dogs that presented with clinical lesions consistent with a superficial pyoderma (papules, pustules, epidermal collarettes, crusts, and ulcers) and/or deep pyoderma (ulcers and draining tracts) were recruited. To be included, dogs had to have predominantly coccoid-shaped bacteria on cytology with Diff-Quik stain^a. Informed consent was obtained from the owners of all the dogs before enrollment into the study. This study was approved by the Iowa State University Institutional Animal Care and Use Committee.

Collection

As seen in Figure 1, the clinical lesions were sampled with a sterile culture swab^b by applying it to the lesion maintaining sterility. The swab was then rinsed in 0.9 mL of sterile phosphate-buffered saline and vortexed for 30 seconds to create a homogeneous solution. Four sterile culture swabs^b were dipped into the homogeneous solution sequentially. The order of culture swab submission to laboratories was determined via computer-generated randomization except for Laboratory 4 always being the last sample. All swabs were submitted as directed to the following laboratories for culture: Laboratory 1 (IDEXX); Laboratory 2 (Zoetis); Laboratory 3 (Antech); Laboratory 4 (Iowa State University Veterinary Diagnostic Laboratory).

View Figure

• Download as Powerpoint
• Download Figure

Data Collection

All data were collected to include the following: submission date, date results finalized, presence of organism detected, specific Staphylococcus spp. identified, quantity of Staphylococcus spp., and antimicrobial susceptibility testing results. Quantity of Staphylococcus spp. from each laboratory were evaluated in the categories listed in Table 1. All laboratories follow CLSI guidelines for antimicrobial susceptibility testing.

TABLE 1 Bacterial Quantity Scores Based on Laboratory Growth

View Fullscreen

Results

Finalized Culture Results

When comparing the number of days for results to be finalized across all laboratories, the data were non-normal in distribution. Unfortunately, Laboratory 2 disposed of four submitted samples without performing culture and susceptibility, which resulted in 20 samples from Laboratory 2 for comparison purposes, whereas the other three laboratories had 24 samples. The results are shown in Table 2 with the median number of days for the results to be available as well as the range. When comparing the number of days for finalization of results, there was a statistically significant difference among the four laboratories (P < .001). Post hoc analysis identified a difference when comparing Laboratory 3 with the others (Laboratory 1 versus Laboratory 3, P = .001; Laboratory 2 versus Laboratory 3, P = .000; Laboratory 3 versus Laboratory 4, P = .005). No difference was identified when comparing Laboratory 1, Laboratory 2, and Laboratory 4 in pairs.

TABLE 2 Laboratory Turnaround Time

View Fullscreen

Staphylococcus spp. Identified

Only 20 samples were evaluated for this purpose as all four laboratories had performed testing on each of these cases. Staphylococcus spp. were identified in 18/20 (90%) samples from Laboratory 1, 19/20 (95%) samples from Laboratory 2, 16/20 (80%) samples from Laboratory 3, and 20/20 (100%) samples from Laboratory 4 (Table 3). In one case, Laboratory 4 was the only facility to isolate a Staphylococcus spp., and in this instance, a few colonies were recovered on culture. Another sample only had a Staphylococcus spp. identified from Laboratory 2 (light growth) and Laboratory 4 (heavy growth). When evaluating for agreement between all four laboratories that a Staphylococcus spp. was present, there was fair agreement (κ = 0.322) based on Fleiss Multirater Kappa.

TABLE 3 Staphylococcus spp. Organisms Detected

View Fullscreen

S pseudintermedius was identified in 16/20 (80%) samples by Laboratory 1, 18/20 (90%) samples by Laboratory 2, 12/20 (60%) samples by Laboratory 3, and 18/20 (90%) samples by Laboratory 4 (Table 3). Other Staphylococcus spp. identified by Laboratory 1 included Staphylococcus intermedius and Staphylococcus schleiferi. None of the samples from Laboratory 1 had more than one Staphylococcus sp. identified. Laboratory 2 identified Staphylococcus felis, S schleiferi, and S aureus as additional Staphylococcus spp. Laboratory 2 did have three samples that identified multiple Staphylococcus spp., including one with S pseudintermedius and S felis, one with S pseudintermedius and S schleiferi, and one with S pseudintermedius and S aureus. Laboratory 3 reported S schleiferi on four samples and coagulase-negative Staphylococcus spp. in two samples. None of the samples from Laboratory 3 reported multiple Staphylococcus spp. Laboratory 4 additionally identified S schleiferi in two samples, with one of them being isolated as the only organism. Four of the samples had reported multiple Staphylococcus spp. on the cultures. These included Staphylococcus lugdenensis, S aureus, Staphylococcus simulans, and coagulase-negative Staphylococcus spp. all in combination with either S pseudintermedius or S schleiferi.

To compare the agreement of the Staphylococcus spp. identified, comparison was evaluated on the 16 samples that identified a Staphylococcus sp. When comparing all four laboratories, there was fair agreement (κ = 0.374) based on Fleiss Multirater Kappa. Laboratory 3 failed to identify the same Staphylococcus sp. compared with the other laboratories, which occurred in 4/16 (25%) samples. Therefore, there was very good agreement when comparing pairs with a Cohen’s Kappa of 1.00 between Laboratory 1 versus Laboratory 2, Laboratory 1 versus Laboratory 4, and Laboratory 2 versus Laboratory 4. There was poor agreement (κ = 0.176) when comparing Laboratory 3 versus Laboratory 1, Laboratory 3 versus Laboratory 2, and Laboratory 3 versus Laboratory 4.

Discussion

Diagnostic turnaround time is a key factor in laboratory selection for bacterial culture and susceptibility. Laboratory 3 was the only one identified as statistically different; it was slower than the other three laboratories, with a turnaround time of 8 days. One factor to consider is that Laboratory 4 was the only laboratory able to be processed within the hospital and did not require shipment to start the culture. Although Laboratories 1–3 required shipment for processing, all of the samples were sent out at the same time, which cannot justify the delay in turnaround time seen with Laboratory 3. Laboratories 1, 2, and 4 resulted in a median time of 6 days or less.

A variety of techniques have been discussed to sample the skin and ears for comparing cultures between laboratories. Some have proposed to sample the same location in sequential order with separate culture swabs, but this technique could potentially have variable amount and type of bacteria obtained. Additionally, studies evaluating culturing of the ears showed variability with placing multiple culture swabs into each ear.^10,11 This could cause variability in the results as the amount of material may not be evenly dispersed among the swabs. Making a homogeneous solution for the samples as done in this study is one way to decrease the variability.

Overall, there was a high detection of Staphylococcus spp. (80–100%) by all laboratories when coccoid-shaped bacteria were seen on cytology. Laboratory 4 was the only facility where samples could be processed onsite on the same day they were acquired. This may be a factor as to why Staphylococcus spp. were detected in 100% of the cases compared with 80–95% from the remaining laboratories. There is a potential that organisms may have become nonviable during transit, leading to lower detection or lower quantity of the organism. Laboratory 4 had statistically higher quantities of Staphylococcus spp. detected, which supports this hypothesis as well. Samples for laboratories 1, 2, and 3 were all shipped to locations between 600 and 650 miles away; therefore, proximity to the reference laboratory may influence results.

S pseudintermedius was the most common isolated species from aerobic culture between the laboratories. When comparing the 16 samples from all four laboratories, Laboratories 1, 2, and 4 had the same results. Laboratory 3 identified a different Staphylococcus sp. in 25% of the cases. The concern for identification of a different Staphylococcus species is whether it is interpreted as a pathogenic organism. Coagulase-negative Staphylococcus spp. are organisms that have previously been considered insignificant. However, newer investigations are shedding light on the pathogenic potential of these organisms in certain circumstances.¹² If susceptibility testing is not performed on these bacteria because of this misconception, clinical improvement may be impacted due to inappropriate antimicrobial selection. Coagulase-negative Staphylococcus spp. were identified from two cases in this study, and susceptibility testing was performed by all laboratories.

Interpretation for methicillin-resistant Staphylococcus spp. had very good agreement when compared among all four laboratories. This is reassuring as methicillin resistance is of high clinical concern when Staphylococcus spp. are encountered in veterinary patients presenting for dermatologic issues. Laboratories 1 and 2 reported oxacillin susceptibility on their antibiotic panel to indicate methicillin resistance. Laboratory 3 did not show oxacillin susceptibility on their antibiotic panel but identified methicillin-resistant Staphylococcus spp. based on oxacillin susceptibility. Laboratory 4 uses oxacillin susceptibility testing along with a commercially available Latex Agglutination Test^d to determine whether PBP2a is present to determine methicillin resistance. This combination of tests has been shown to have high positive predictive values (100%) and negative predictive values (99%).¹³

Good to very good agreement was noted for all antimicrobials among all four laboratories. Cumulatively, there was more than 94% agreement on all organisms and antimicrobials. This is in contrast to studies evaluating susceptibility profiles for Pseudomonas spp. in the ears of dogs with agreement only noted 25–32% of the time.^11,14 BMD is considered the reference standard by CLSI and EUCAST for antimicrobial susceptibility. Previous studies have shown variable MIC values depending on the commercial test methods when compared with BMD.⁴ Investigators in that study observed lower MIC values with the Vitek system, whereas the Sensititre system was shown to have higher MIC values for some antimicrobials. These results were identified with β-lactams when testing S pneumoniae but not confirmed in our study.⁴ The Vitek system was used by Laboratories 1 and 2, and the Sensititre system was used by Laboratories 3 and 4. The low correlation with clindamycin (only good agreement) is somewhat concerning as it is considered a tier 1 antibiotic for treatment of superficial bacterial folliculitis.²

One of the limitations of this study is the sample size. A more robust multi-institutional study would have been ideal to determine whether a significant difference was missed in the current study. We felt it was ideal to test some of the most commonly used commercial veterinary laboratories to make it clinically applicable. Some of the smaller and regionally exclusive veterinary laboratories would have provided another option to determine whether the size of the laboratory influences the results. Future studies could include multiple university veterinary diagnostic laboratories or limiting the study to laboratories accredited through the American Association of Veterinary Laboratory Diagnosticians.

Conclusion

There was mild variability identified among the four laboratories in our study. Laboratory 3 tended to have a slower turnaround time, identified Staphylococcus spp. at a lower rate, and identified a different Staphylococcus sp. in 25% of cases compared with the other three laboratories. Antimicrobial susceptibility testing had good or very good agreement across all four laboratories with very good agreement on methicillin resistance. A combination of identification of the offending bacterial organism as well as antimicrobial susceptibility is of the utmost importance to obtain clinical resolution in cases of pyoderma. Therefore, finding a laboratory that you trust for reliable results is important.