Investigation of Reported Resistance of Streptococcus equi ssp. zooepidemicus to Trimethoprim-Sulfa in Horses

Equine veterinarians have worked with laboratory staffs to investigate a marked change in reported susceptibility of Streptococcus zooepidemicus to trimethoprim-sulfa. This study shows that deviation from the National Committee for Clinical Laboratory Standards (NCCLS) guidelines can result in false reporting of resistance of S. zooepidemicus to trimethoprim-sulfa. False reporting of resistance by the laboratory can affect treatment decisions made by equine veterinarians; thus, it is important that the laboratory comply with the recommended guidelines for antimicrobial susceptibility testing.

*Presenting author.

1. Introduction

Results of laboratory testing of susceptibility of bacteria to selected antimicrobials is often incorporated into the decision-making process regarding the best antimicrobial to use to treat an equine patient with an infection. Further considerations in choosing an antimicrobial include associated cost, ease of administration, potential adverse side effects, and pharmacokinetic characteristics.

Two of the more common antimicrobial susceptibility testing (AST) methods used in the laboratory are the Kirby-Bauer disk diffusion method and the broth microdilution method. The Kirby-Bauer disk diffusion method is most commonly used in testing equine isolates based on a survey of veterinary diagnostic laboratories [1]. Guidelines for AST are published by the National Committee for Clinical Laboratory Standards (NCCLS) and contain details regarding the recommended media and organisms for quality assurance testing [2].

Trimethoprim-sulfamethoxazole (SXT) and trimethoprim-sulfadiazine are antimicrobial drug combinations (potentiated sulfonamides) widely used in veterinary medicine to treat a range of bacterial infections. Potentiated sulfonamides are a particularly popular choice of antimicrobial drug for use in equine patients for a variety of reasons: the bactericidal effect against many gram-positive and gram-negative bacteria, the oral route of administration, the relative safety, and the low cost. Potentiated sulfonamides frequently are selected to treat respiratory infections in equine patients; this decision is based on knowledge of likely microorganisms and their predicted susceptibilities, and it is supported by bacterial culture and susceptibility results [3].

Streptococcus equi ssp. zooepidemicus, a Lancefield group C, β -hemolytic cocci, is reported to be the most common microorganism associated with lower respiratory tract disease in horses [3-6]. Historically, most β -hemolytic streptococci isolated from equine patients have been considered predictably susceptible to β -lactam and potentiated sulfonamide antimicrobial agents [3,7]. Consequently, potentiated sulfonamides frequently are part of the treatment of such infections in horses.

The purpose of this retrospective study was to determine if there was increasing resistance to SXT in S. zooepidemicus isolated from horses; subsequently, the impact of bacterial susceptibility testing methods on the reported susceptibility of S. zooepidemicus to SXT was investigated.

2. Materials and Methods

Part 1

A retrospective study was conducted through the Colorado State University James L. Voss Veterinary Teaching Hospital (CSU-VDL) medical records database, identifying a sample size of 50 S. zooepidemicus isolates from the respiratory tract of equine patients during each of two separate time periods: 1987-1990 and 1997-2001. The number of cases of S. zooepidemicus reported as resistant to SXT from each time period were compared and statistically analyzed using a cross-sectional study design.

Part 2

Further investigation involved comparing the susceptibility of a separate group of 41 S. zooepidemicus isolates of equine origin. Both the Kirby-Bauer test, as it was being performed at the CSU-VDL before this investigation, and the broth microdilution method were used. The source of the equine isolates for this part of the study were those recovered from samples submitted to the CSU-VDL [a].The technique used for antimicrobial susceptibility testing performed at the Animal Veterinary Diagnostic Laboratory in Purdue was the Sensititre/Trek broth dilution method; this method added laked horse blood to the Mueller-Hinton broth medium for testing with SXT, according to the recommendations of the NCCLS [2]. Recommended quality control data were obtained to ensure the validity of the tests [8].

3. Results

Part 1

In the time period of 1987-1990, 4 of 50 (8%) S. zooepidemicus isolates were reported in the medical record to be resistant to SXT, using the quantitative microbroth dilution (Trek) system.

In the time period 1997-2001, 21 of 50 (42%) S. zooepidemicus isolates were reported in the medical record as resistant to SXT by the laboratory, using Mueller-Hinton broth with 5% sheep red blood cells as the culture medium [b]. This difference was statistically significant based on an odds ratio (OR) of 8.3 and p = 0.0001.

Part 2

Based on results of part 1, further investigation into the method of testing of β -hemolytic Streptococcus spp. for susceptibility to SXT in the more recent time period was undertaken. The culture medium (Mueller-Hinton medium with 5% sheep red blood cells) used for the susceptibility testing of S. zooepidemicus to SXT using the Kirby-Bauer disk diffusion technique was determined to be unsuitable based on recommended quality-control procedures. High levels of para-aminobenzoic acid (PABA) and its analogues are present in whole sheep blood and cause reduced antimicrobial activity of trimethoprim and sulfamethoxazole; this results in smaller inhibition zones and interzonal colony growth. Thus, bacterial isolates were being read as either susceptible or resistant based on indistinct zones of growth inhibition, yielding false-resistance reports.

The problem with the testing method for β -hemolytic Streptococcus spp. against SXT was not detected until after the widespread use of the Kirby-Bauer testing method. However, results indicate that the increase in false-resistance reports of S. zooepidemicus to SXT occurred after July 1999, long after Kirby-Bauer test was introduced into the laboratory. The reason for this change was not definitively determined, but it is possible it was caused by a change in the agar media.

Any change in agar media would have been detected if the appropriate quality-control bacteria had been used. However, because the agar media was not being tested with the NCCLS-recommended organism (Enterococcus faecalis American Type Culture Collection [ATCC] [c] 33186 or ATCC 29212) but instead with Streptococcus pneumoniae (ATCC 49619), the problem with the SXT-susceptibility testing was not detected until the clinical retrospective study (e.g., part 1 of the investigation) was conducted. The seven equine S. zooepidemicus isolates available at the time of the investigation were retested using the disc-diffusion technique with alternative growth medium. Results classified all seven isolates, which had originally been classified as resistant, as susceptible to SXT.

In the pilot antimicrobial resistance study, results from the testing of a separate group of 41 equine-origin S. zooepidemicus isolates showed that 15 of 41 isolates (37%) were reported as resistant to SXT by the CSU-VDL when using the original Kirby-Bauer test and the Mueller-Hinton medium mixed with 5% sheep red blood cells. In comparison, 100% of the same 41 S. zooepidemicus isolates that were tested using broth micro-dilution methods were reported as susceptible to SXT.

Through steps undertaken in part 2 of this study, it became clear that much of the resistance of S. zooepidemicus to SXT that was reported by the CSU-VTH laboratory was the result of a flawed testing method.

4. Discussion

Diseases caused by Streptococcus spp. infection are common in veterinary medicine. It has been suggested that susceptibility testing is not always necessary for clinical isolates, such as β -hemolytic Streptococcus spp., because some authors have reported that the majority of equine β -hemolytic Streptococcus isolates are predictably susceptible to certain antimicrobials [3]. However, at the CSU-VDL, antimicrobial susceptibility is routinely performed to determine the susceptibility of all β -hemolytic Streptococcus spp. isolates of equine origin to a panel of antimicrobials. The increase in SXT resistance to the common equine pathogen S. zooepidemicus, as reported by the CSU-VDL, warranted further study.

Our investigation revealed a problem in susceptibility testing of β -hemolytic Streptococcus spp. to SXT. Prescott [9] describes problems in laboratory testing of Streptococcus spp. , as well as a number of other microorganisms, to diaminopyrimidine-sulfonamide drug combinations that are caused by PABA or thymidine in whole blood-based media. PABA and its analogs show antagonism to sulfonamides and trimethoprim. This phenomenon is well recognized and described in laboratory manuals [10,11,12]. If AST guidelines are not followed in detail, including the use of the appropriate quality-control microorganisms, it can impact susceptibility results by causing zones of growth inhibition to be smaller, less distinct, or even, result in no zone at all [8]. The published NCCLS guidelines[10] recommend using media containing lysed (laked) horse blood or thymidine phosphorylase for testing of β -hemolytic Streptococcus spp. with potentiated sulfonamides. It is believed that laked horse blood is likely acceptable because of its high concentration of thymidine phosphorylase, an enzyme that deactivates thymidine [9]. These additions can improve the clarity of zones and therefore, the reliability of trimethoprim and sulfamethoxazole susceptibility testing of Streptococcus spp [8]. The PABA and thymine/thymidine content of Mueller-Hinton medium and broth are reduced to a minimum [9], and results with unmodified Mueller-Hinton have been generally satisfactory for the testing of Streptococcus spp. to sulfonamide and trimethoprim. However, Streptococci spp. may require enrichment media supplemented with blood or blood products to grow at an adequate rate [9]. Although antimicrobial antagonism by blood products exists in vitro, it apparently is not present in vivo. This may be explained by the presence of low levels of thymidine and thymine in blood that do not interfere with in vivo antibacterial activity during treatment [9].

Before this study, a commercial source for laked horse blood-supplemented agar was unavailable to the CSU-VDL. Subsequently, a suitable commercial manufacturer for laked horse blood in Mueller-Hinton agar was identified. Currently, testing β -hemolytic Streptococcus spp. with trimethoprim and sulfamethoxazole is performed using Mueller-Hinton media supplemented with 5% laked horse blood[d] as recommended and approved by the NCCLS [10]. Quality control has also been enhanced with adherence to the NCCLS guidelines. Since making these modifications, the CSU-VDL has not recognized any previously questionable zones of inhibition on susceptibility plates when testing susceptibility of β -hemolytic Streptococcus spp. of equine origin to SXT.

There is recent evidence suggesting that a number of microbiology laboratories are failing to comply with NCCLS recommendations for quality control of culture media [13]. A survey of 124 laboratories performing testing of human samples in Ontario showed that few laboratories actually followed recommended quality control protocols [13]. Major compliance failures included failure to perform quality control at all and inappropriate choice of bacterial strains and inoculum used for quality control [13]. Although the findings of this survey are not specific for the susceptibility of S. zooepidemicus to SXT as investigated in this report, it may suggest that not all laboratories comply fully with recommended practices

Considering the findings from the above-mentioned surveys and the information learned from our retrospective study and subsequent investigation, further assessment and more rigid control of standard laboratory procedures are warranted. Although a number of publications clearly describe the laboratory techniques that should be employed to obtain the most accurate information regarding antimicrobial susceptibility testing, these guidelines may not be followed in adequate detail to lead to valid testing. Consequently, laboratories may be reporting inaccurate information that can significantly affect antimicrobial drug choice made by clinicians in the treatment of clinical cases. These problems may easily go unrecognized, unless trends in antimicrobial susceptibility patterns are critically evaluated for trends over time and detectable changes that may occur from one time period to another.

The effectiveness of antimicrobials against a specific organism is best determined by a treatment-outcome investigation of the infection in a specific site. There has been one study of the effectiveness of trimethoprim-sulfa in killing S. zooepidemicus in an in vivo experimental model in ponies [14]. The trimethoprim-sulfadiazine reduced the number of bacteria, but it did not eliminate the infection [14]. The authors speculated the reason for treatment failure was the lack of efficacy of the trimethoprim in the presence of pus [14]. Thus, the equine clinician must decide the most appropriate antimicrobial treatment based on several considerations, including the results of laboratory AST and the likely effectiveness of the antimicrobial at the site of action.

This study was funded in part by a grant from the United States Department of Agriculture and the Cooperative State Research Education and Extension Services for the Colorado State University Program for Economically Important Infectious Animal Diseases. We acknowledge Gayle Thompson, Barbara Traut, Jean Muirhead, and Denise Bolte (laboratory technicians at Colorado State University Veterinary Diagnostic Laboratory) for their technical assistance.

*A short communication of part of this material was published in the Journal of Veterinary Diagnostic Investigation. September 2005; 17(5):483-486.

Footnotes

[a] American Veterinary Medical Association (AVMA) Steering Committee on Antimicrobial Resistance (SCAR), Surveillance System of Antimicrobial Susceptibility for Animal Pathogens project. Unpublished data.
[b] BBL, Becton Dickinson Microbiology Systems, Sparks, MD 21152.
[c] American Type Culture Collection, Manassas, VA 02108.
[d] Hardy Diagnostics, Santa Maria, CA 93455.