Concentrating on students to improve antimicrobial stewardship

In their recent letter,

Sullivan et al¹ suggest that further attention should be focused on educating veterinary students on antimicrobial resistance and the principles of antimicrobial stewardship. However, in my more than 50-year career, I have seen little improvement in antimicrobial use, and I doubt changing the training that veterinary students receive would have a substantial impact. This would especially seem to be the case because veterinarians learn most of their clinical skills after leaving veterinary school.

When it comes to anti-microbial use, many veterinary practitioners default to what they have done in the past and prescribe antimicrobials without further evaluation. In at least some instances, doing nothing and letting nature take its course would yield the same result. Our culture is such that we often have to do something even though it is unlikely to make a substantial difference in the outcome. Still, veterinarians should make serious efforts to use antimicrobials only when necessary. This starts with following standard protocols and performing the required follow-up to verify efficacy.

Importantly, veterinarians must realize that they are not the only ones who have an influence on when and what antimicrobials will be used in any particular disease scenario. Improving antimicrobial stewardship will require discussions with those outside the veterinary profession. There is more to be learned about this issue than we presently know, and solving the problem of antimicrobial resistance will take expertise from more than medical professionals.

Economic incentives will likely also be necessary to improve antimicrobial stewardship, particularly in the area of production animal medicine. Producers will need to be paid to invest in production improvements and alternatives to antimicrobials. In the end, consumers will bear the costs. Still, current concerns that the use of antimicrobials in animal agriculture will have a negative effect on human health are based on equivocal research findings, and more information is needed on the costs and benefits of administering antimicrobials at low doses for long periods versus at high doses for short periods.

Additional information is also needed on potential alternatives to antimicrobials, especially technologies based on bacteriophages (viruses that infect bacteria). Phages are ubiquitous and essential to life on Earth, and phage-based technologies have been used successfully for many generations. Still, our knowledge of what goes on at the cell level is very basic, and we have little knowledge of what really happens as phages target and kill bacteria.

We need to think outside the box and not accept that antimicrobial use as we know it today is correct. There are many alternatives that we have not seriously explored, but veterinary student teaching will likely not take the lead in these areas.

Joseph Butterweck, dvm

Fresno, Calif

More on declawing of cats

The recent letter by Dr. Nolan Beech¹ and the AVMA response² got me thinking and reading. I encourage AVMA members to read the literature review³ mentioned in the AVMA response and to decide, given the information therein, what they believe is the best policy for their own patients.

As I type this, my 10-year-old declawed indoor cat, Nina, is lying on my lap happily kneading my leg. Nina has never scratched anyone and has damaged no furnishings. She is active and playful, a veteran pouncer, and a regular pseudoscratcher of the edges of my couch. I recognize that the AVMA invested a lot of thought when it amended its policy on declawing.⁴ However, I wonder how much consideration was given to the possible adverse effects on cats themselves if declawing were to become less available because of the policy changes. For example, the literature review states that “in a 1991 survey of Ontario veterinarians, respondents indicated that approximately 50% of their clients would no longer own their cat if it had not been declawed.” It seems clear that many possible cat owners either cannot, fail to, or simply choose not to make the effort to “explore and implement nonsurgical alternatives,” as suggested in the literature review. Thus, if effectively promoted and implemented, the AVMA policy will inevitably reduce the number of cat-owning households. Notwithstanding the impact on veterinary practice activity, the result will be that many cats will not be placed in homes and will be either killed or left to a short feral life. What is the moral implication of this unintended but very real outcome? Nina cannot speak for herself, but I feel certain she would choose her pampered, playful, declawed life instead of death.

John Fetrow, vmd, mba, dsc

Langley, Wash

Developing antibiograms

We read the recent study by Frey and Jacob¹ and the accompanying commentary² and applaud the attention they bring to developing antibiograms for companion animal practice, especially given the positive effects this could have on antimicrobial stewardship. However, on behalf of the Clinical and Laboratory Standards Institute (CLSI) Subcommittee on Veterinary Antimicrobial Susceptibility Testing, we would like to address several concerns about the authors’ methods and interpretation.

In particular, some of the CLSI guidelines for presentation of antibiograms³ were not followed by Frey and Jacob. For example, the CLSI guidelines recommend incorporating data for a minimum of 30 isolates of each bacterial species, but the authors elected to use a cutoff of 15 isolates. The recommendation to include a minimum of 30 isolates per bacterial species was made to reduce the possibility of missing resistant isolates and to minimize the influence of clonal isolates. Table 4 indicates there were only 10 isolates of Staphylococcus pseudintermedius and that all 10 were susceptible to oxacillin. However, if a larger population of isolates had been included, it is extremely unlikely that 100% of S pseudintermedius strains would have been classified as susceptible. When a minimum of 30 isolates is not available, we would propose expanding the time frame beyond a calendar year, rather than reducing the required number of isolates.

Table 5 reports that 88% of Enterobacterales (formerly Enterobacteriaceae) strains were susceptible to amoxicillin–clavulanic acid, 38% were susceptible to cephalexin, 81% were susceptible to amoxicillin, and 58% were susceptible to doxycycline. However, all wild-type Enterobacterales typically test resistant to these agents, and comparison of reported median minimum inhibitory concentrations to approved breakpoints⁴ suggests that the reported values overestimate the percentage susceptible. Additionally, ceftiofur and cefovecin, which were included in the table, should not be reported for Enterobacterales isolates from sites other than urine.

Table 6 reports that 83% and 95% of Escherichia coli isolates from private practices and an academic practice, respectively, were susceptible to doxycycline. However, the reported median minimum inhibitory concentration was 1 µg/mL, whereas the clinical breakpoint doxycycline concentration in dogs is much lower (0.12 µg/mL), and the CLSI recommends that all wild-type E coli isolates be considered resistant.⁵

Finally, CLSI antimicrobial breakpoint concentrations for urine isolates should not be applied to isolates from the bladder wall. Thus, interpretations presented in Table 4 may not be applicable to uncomplicated lower urinary tract infections. Also, 13% and 36% of the private practice and academic practice patients from which urine and bladder wall isolates were obtained were cats, but urine-specific breakpoints are not available for interpretation of feline samples, and inclusion of feline isolates may have skewed the values in Table 4.

Because the primary objective of an antibiogram is to guide clinicians, on the basis of susceptibility of isolates from similar cases, in the empirical selection of antimicrobial treatment, the results provided by Frey and Jacob may be misleading for veterinary practitioners. For example, if a clinician were to select amoxicillin for treatment of a skin infection on the basis of the fact that 81% of Enterobacterales were susceptible to amoxicillin, the likely result would be treatment failure, because pharmacokinetic data indicate there is no chance of reaching the reported minimum inhibitory concentration of 4 µg/mL with amoxicillin in dogs. We are concerned that the report overestimates the percentage of isolates susceptible to various antimicrobials, which could lead clinicians to expect a particular antimicrobial to work when, in reality, it is unlikely to provide any true clinical benefit. We suggest that the interpretations presented in the report be revised to avoid a potential negative impact on patient care.

Brian V. Lubbers, dvm, phd, dacvcp

Chairholder

Dubraska V. Diaz-Campos, dvm, phd

Vice Chairholder

Mark G. Papich, dvm, ms, dacvcp

Past Chairholder

Jeffrey L. Watts, phd, m(ascp), rm(nrcm)

Past Chairholder

Subcommittee on Veterinary

Antimicrobial Susceptibility Testing

Clinical and Laboratory Standards Institute

Malvern, Pa

The authors respond

We thank the Clinical and Laboratory Standards Institute (CLSI) Subcommittee on Veterinary Antimicrobial Susceptibility Testing for highlighting the many challenges of creating useful and representative antibiograms to enhance antimicrobial stewardship. In sharing our methods for developing antibiograms in a private practice setting, our goal was to follow the principles of the CLSI M39-A4 document¹ or explain why we could not. We aimed to highlight the need for bacterial culture and antimicrobial susceptibility testing (AST) for decision-making and to demonstrate a method to make AST data analysis approachable.

Because veterinarians may be unfamiliar with antibiograms, we believed it was necessary to include examples. To guard against potential misuse, we labeled the tables as representative and emphasized that they should not be used to make clinical decisions. We also organized the tables to be in line with relevant guidelines from the International Society for Companion Animal Infectious Diseases^2,3 to illustrate how these guidelines could be integrated into the antibiogram to enhance decision-making, rather than to be prescriptive about drug choice.

The antimicrobial drugs and interpretations included were those reported by the reference laboratory. Most practitioners are unaware of breakpoint changes or reasons for variations in antimicrobial drug reporting within or between laboratories and must make clinical decisions on the basis of interpretations (ie, susceptible, intermediate, or resistant) reported by their laboratory. We added median minimum inhibitory concentrations to allow readers to reinterpret our findings on the basis of updated breakpoints, and we noted those breakpoints changed in 2018.⁴ We, and importantly practicing veterinarians, could not know if or when these updated breakpoints were used by the laboratory.

In addition, we recommended that practitioners compare antimicrobial drugs and breakpoints on their AST reports with published standards to assess alignment.^1,4,5,6 Because updating laboratory breakpoints may require time, changes will not be immediately apparent to practicing veterinarians. We encourage the CLSI to continue their efforts to engage with diagnostic laboratories to promote the adoption of updated breakpoints and reporting guidelines.

Because we limited ourselves to AST data from a single year, it was challenging to meet CLSI's recommendation of 30 isolates. This reinforces the need for veterinarians to perform more culture and AST. We acknowledged the benefits of a larger sample, and we agree that expanding the data collection time frame would be another solution. We encourage veterinarians to consider this option.

The authors noted that the clinical breakpoint of doxycycline in dogs is 0.12 μg/mL; however, from the available AST reports, it appeared that minimum inhibitory concentrations of ≤ 0.5, 1.0, 2.0, and 4.0 μg/mL were interpreted as susceptible, a concentration of 8.0 μg/mL was interpreted as intermediate, and concentrations ≥ 16 μg/mL were interpreted as resistant. We analyzed the breakpoints and interpretations provided by the laboratory just as practicing veterinarians would. In addition, our data contained only a single sample from the urinary bladder wall, which was reported to have no growth and was not included in the final cumulative antibiogram. We agree that only samples from urine should be used, in keeping with the CLSI document M39-A4.¹

Our report highlights the challenge in obtaining a statistically valid number of isolates and emphasizes the importance of the volume and currency of AST reports. We appreciate that CLSI has made some veterinary resources freely available and suggest veterinarians use these in support of antimicrobial stewardship.^4,6

Erin Frey, dvm, mph

Megan Jacob, phd

North Carolina State University

Raleigh, NC