Introduction
Rhodococcus equi is a gram-positive, facultative, intracellular pathogen that causes severe pyogranulomatous pneumonia in many animal species and humans,1,2 most commonly in young horses that range in age from 3 weeks to 5 months.3,4 The disease adversely impacts farms where this disease is endemic because of costs for prolonged treatment of affected foals, substantial morbidity and mortality rates, and decreased racing performance of foals that recover.5 To detect this disease sooner and reduce its impacts, many horse-breeding farms in the United States have implemented sequential TUS followed by prophylactic antimicrobial treatment of foals identified to have pulmonary lesions consistent with subclinical infection with R equi.6,7 This practice can decrease mortality rates at farms with endemic disease6 but has unclear efficacy because most subclinically affected foals recover spontaneously.8 Moreover, the efficacy of antimicrobial treatment can be comparable to the use of a placebo in foals with subclinical pneumonia.9
The synergistic combination of a macrolide (most often erythromycin, azithromycin, or clarithromycin) with rifampicin has been a standard treatment for R equi infections since the 1980s.10 Before around 2001 when TUS was implemented to detect subclinical pneumonia on farms where R equi was endemic, reports of R equi isolates resistant to macrolides or rifampicin were rare.11 Recently, however, there has been an alarming increase in antimicrobial-resistant R equi isolates cultured from clinical and environmental samples.12,13,14,15 For example, antimicrobial-resistant R equi isolates were cultured from soil samples obtained from 76 of the 100 horse-breeding farms tested in central Kentucky during 2018, and greater use of macrolides and rifampin to treat foals with pneumonia was associated with increased odds of isolating MRRE from soil samples.14 To further evaluate that positive association, our objectives with the prospective observational study reported here were to compare soil concentrations of MRRE at horse-breeding farms that used the combination of TUS to identify foals with subclinical pneumonia and subsequent administration of macrolides and rifampin to affected foals (TUS farms) versus soil concentrations of MRRE at farms that did not use the combination of TUS and subsequent antimicrobial treatment of foals with subclinical pneumonia (non-TUS farms), determine whether the combined use of TUS and antimicrobial treatment of foals with subclinical pneumonia was associated with soil concentration of MRRE, and assess whether there were temporal effects on soil concentrations of MRRE during the foaling season. We hypothesized that soil concentrations of MRRE would be higher for TUS farms versus non-TUS farms.
Materials and Methods
Farm selection
Twenty horse-breeding farms in central Kentucky were selected from 72 farms that had been enrolled in a prior study14 and had complete records available. A sampling function in commercial softwarea was used to randomly select 10 TUS farms from those that reported having used TUS with subsequent prophylactic antimicrobial treatment of subclinical pneumonia attributed to R equi for a minimum of 5 consecutive years (n = 53) and 10 non-TUS farms from those that reported having never used TUS (n = 19). Study protocols were reviewed and approved by the Clinical Research Committee of the University of Georgia College of Veterinary Medicine, and informed client consent was obtained from all participating farms before enrollment.
Soil sample collection
At each farm, 3 paddocks that were most frequently used by mares and foals were identified by each farm manager for sampling during the 2019 foaling season, and samples were collected at each sampling time from where the manager indicated mares and foals tended to congregate. Three soil samples were collected from 3 separate sites in each of the 3 individual paddocks (9 samples/farm) for each round of sampling; to the extent possible, the same sites were sampled for each round. At each farm, a round of sampling was performed during the last 2 weeks of January, March, May, and July 2019, with all samples collected between 8:00 am and 12:00 pm.
Sample processing
Collected soil samples were stored and processed as previously described.13,14 Briefly, for each soil sample, 1 g of soil was quantitively cultured on separate plates containing NANAT media (selective for R equi), modified NANAT media with added erythromycin (8 μg/mL; selective for macrolide-resistant R equi), or modified NANAT media with added rifampicin (50 μg/mL; selective for rifampicin-resistant R equi). Colony identification and virulence status were determined with PCR assay amplification of the choE and vapA genes as previously described.16,17 To confirm antimicrobial resistance, minimum inhibitory concentrations of azithromycin, clarithromycin, erythromycin, and rifampicin were determined through the use of antimicrobial concentration gradient test strips,b according to the manufacturer's recommendations and guidelines established by the Clinical and Laboratory Standards Institute.18,19 Concentrations of antimicrobial agents tested were 2-fold dilutions between 256 and 0.016 mg/L for all macrolides and between 32 and 0.002 mg/L for rifampicin. Control isolates tested in parallel and on each test occasion for all methods were Staphylococcus aureus ATCC 29213 and Enterococcus faecalis ATCC 29212.20 Ultimately, the concentration of MRRE in each sample was quantified by enumeration of CFUs per gram of paddock soil from a dilutional series on selective media. A mean soil concentration of MRRE was calculated for each of the 3 paddocks sampled each time on each farm, and these 3 mean soil concentrations of MRRE were summed; the unit for data analysis was the sum of the mean soil concentrations of MRRE per sample month per farm.
Survey data collection
In addition to sampling soil from participating farms, information regarding each farm's 2019 foaling season, including the number of foals born, number of foals identified with pneumonia attributed to R equi, number of foals treated with macrolides alone or in combination with rifampicin, and whether the farm used TUS for early detection of R equi pneumonia, was collected through the use of a questionnaire (Supplementary Appendix S1, available at: avmajournals.avma.org/doi/suppl/10.2460/javma.258.6.648) that was completed by each participating farm. A hard copy of the questionnaire was provided to the farms at the end of the sample collection process (between July 1 and 31, 2019), and the last completed questionnaire was received August 31, 2019.
Statistical analysis
Sample size calculations were performed on the basis of > 95% power, 10 simulations, α level of 0.05, cluster size of 3 (paddocks per farm), 20 clusters (farms), between-cluster variance of 0.2, an effect size of log(0.75), 4 sampling periods (months), period effect assumed to be constant at log(0.5), farm modeled as a random effect with paddock nested within farm, and 12 at-risk parameters. The value of cluster within-farm variance of effects and the effects of effect size and period effect were based on the judgment by the investigators that these were clinically relevant magnitudes. Calculations were made with available software.a,c,d
Statistical analysis was performed with available software,c and values of P < 0.05 were regarded as significant. Results were analyzed by use of linear mixed-effects modeling to account for the repeated measures within time and farm and by use of a functione that fits a nonlinear mixed-effects model. For linear mixed-effects modeling, the dependent variable was the sum of the 3 mean paddock soil concentrations of MRRE per sample month per farm (log10-transformed CFUs/g), farm was modeled as a random effect, and farm group (TUS vs non-TUS farms), month, and their interaction were modeled as fixed categorical effects. Model fit was assessed by inspection of diagnostic residual plots. Post hoc pairwise testing of differences among months and groups was performed with simultaneous inferencef and the Tukey method. The Wilcoxon rank sum test was used to compare between groups the numbers of foals born and the proportions of foals treated.
Results
All 20 selected farms completed the questionnaire and were sampled as planned. Mares and foals were present in the paddocks at all sampling times. Overall, there were 720 soil samples evaluated (360 from each group [10 TUS farms and 10 non-TUS farms]), and isolates of MRRE were recovered from ≥ 1 soil sample from 19 of the 20 (95%) participating farms.
From January to July 2019, the median number of foals born was 37 (range, 5 to 82 foals) for all 20 participating farms, 47.5 (range, 12 to 82 foals) for TUS farms, and 31.5 (range, 5 to 68 foals) for non-TUS farms. The median number of foals treated with antimicrobials was 2.5 (range, 0 to 20 foals) for all 20 participating farms, 5 (range, 0 to 9 foals) for TUS farms, and 1 (range, 0 to 20 foals) for non-TUS farms. Overall, the median percentage of foals treated with macrolides alone or in combination with rifampicin was 5% (range, 0% [0/56] to 50% [6/12]). The median percentage of foals treated with macrolides alone or in combination with rifampicin was significantly (P = 0.044) higher for TUS farms (9%; range, 0% [0/56] to 50% [6/12]), compared with non-TUS farms (4%; range, 0% [0/15] to 33% [20/60]).
Findings from linear mixed-effects modeling indicated that the overall sum of the 3 mean paddock soil concentrations of MRRE per sample month per farm (log10-transformed CFUs/g) across all months sampled was significantly (P = 0.013) greater and approximately 11-fold higher for TUS farms (8.85 log10-transformed CFUs/g) than for non-TUS farms (7.37 log10-transformed CFUs/g; Figure 1). When effects of sample period were considered, differences were detected on the basis of sample month and farm group (TUS farms vs non-TUS farms; Figure 2). For non-TUS farms, the sum of the mean soil concentrations of MRRE was significantly (P < 0.001) higher in March and May (6.35 and 5.34 log10-transformed CFUs/g, respectively), compared with July (0 log10-transformed CFUs/g). For TUS farms, the sum of the mean soil concentrations of MRRE did not differ significantly among sample months (5.44, 7.39, 6.68, and 7.83 log10-transformed CFUs/g for January, March, May, and July, respectively). During January and July, the sum of the mean soil concentrations of MRRE was significantly (P < 0.05) higher for TUS farms versus non-TUS farms.
Box-and-whisker plots of the sum of the mean soil concentrations of MRRE cultured per sample month (January, March, May, and July) per farm during the 2019 foaling season for TUS farms (light gray; n = 10) versus non-TUS farms (dark gray; 10). For each plot, the box represents the interquartile (25th to 75th percentile) range, the horizontal line in the box represents the median, whiskers represent a multiple of 1.75 of the respective quartile, and each dot represents a farm's sum of the mean soil concentrations of MRRE cultured across all sample months. The sum of the mean soil concentrations of MRRE per farm across all sample months was significantly (P = 0.013) greater for TUS farms (8.85 log10-transformed CFUs/g) versus non-TUS farms (7.37 log10-transformed CFUs/g).
Citation: Journal of the American Veterinary Medical Association 258, 6; 10.2460/javma.258.6.648
Box-and-whisker plots of the sum of the mean soil concentrations of MRRE for the TUS farms (light gray) and non-TUS farms (dark gray) described in Figure 1 considered by sample month (January, March, May, and July). The sum of the mean soil concentrations of MRRE did not differ substantially among sample months for TUS farms but was significantly (P < 0.001) higher in March and May (6.35 and 5.34 log10-transformed CFUs/g, respectively), compared with July (0 log10-transformed CFUs/g), for non-TUS farms. In each plot, each dot represents a farm's sum of the mean soil concentrations of MRRE for that sample month. See Figure 1 for the remainder of the key.
Citation: Journal of the American Veterinary Medical Association 258, 6; 10.2460/javma.258.6.648
Discussion
Results of the present study indicated that farms that used (vs did not use) the combination of TUS to detect subclinical pneumonia followed by antimicrobial treatment of subclinically affected foals had significantly higher soil concentrations of MRRE. This finding supported our hypothesis that soil concentrations of MRRE would be higher for TUS farms versus non-TUS farms. We were motivated to test this hypothesis by findings from previous observational epidemiological studies11,12,13,14 that suggest a temporal link between the emergence of MRRE in clinical and environmental samples from horse farms and the prophylactic administration of a macrolide, alone or in combination with rifampicin, to foals with subclinical respiratory disease identified by the use of TUS. In the present study, the median percentage of foals treated with macrolides, alone or in combination with rifampicin, was substantially higher for TUS farms (9%; range, 0% [0/56] to 50% [6/12]) versus non-TUS farms (4%; range, 0% [0/15] to 33% [20/60]). Because intestinal absorption of orally administered macrolides is incomplete and because of enterohepatic circulation, macrolides and active metabolites of macrolides are excreted in the feces of treated animals.21 It is plausible that greater use of antimicrobials at farms that perform TUS for controlling R equi pneumonia in foals results in selective pressure for MRRE in foals and their environments. Given our evidence of an association between the combined use of TUS and subsequent prophylactic administration of macrolides and rifampin with higher concentrations of MRRE in the soil of farms tested in the present study and evidence that prophylactic treatment of foals with subclinical pneumonia is not always more effective than treatment with a placebo,9,22 the risks and benefits of the practice of prophylactic treatment of foals with pulmonary lesions identified by TUS must be carefully weighed by veterinarians and horse farm managers.
The proportion of foals treated with macrolides at TUS farms and non-TUS farms was lower than expected on the basis of previous studies.12,14,23,24 The reason for this was unknown, and we (NMS and MEG) verified these findings with the staff and veterinarians for these farms. One possible explanation for this finding was that farms in central Kentucky became more prudent about antimicrobial use in light of a recent study14 in which MRRE isolates were cultured from the soil of 76 of 100 farms in the region. In addition, the 20 farms included in the present study were selected from the farms that participated in that previous study.14 Thus, although it was not surprising that MRRE isolates were cultured from soil samples obtained from 19 of 20 (95%) farms in the present study, this finding underscored the high prevalence of MRRE in the environment at horse-breeding farms in this major horse-breeding region.11,12,14
Our findings indicated that the environmental burden of MRRE markedly varied across months of the foaling season for non-TUS farms (highest in March and May) but did not vary substantially for TUS farms. The reasons for this temporal pattern for non-TUS farms were unknown. One possibility was that the MRRE were outcompeted by fitter, susceptible strains of R equi with reduced selective pressure from exposure to administered and excreted antimicrobial substances.25,26 Encouragingly, this finding suggested that reduced use of antimicrobials could diminish or even deplete MRRE from the environment at horse-breeding farms. Alternatively, MRRE might undergo genetic adaptation to improve fitness and persist over time, even when selective pressure from antimicrobials is decreased.27 Further ecological and epidemiological studies of the distribution, persistence, and clonality of MRRE over time in the environment at horse farms in central Kentucky and elsewhere are thus warranted.
As for any observational study, the present study had limitations. First, despite being chosen on the basis of a priori calculations, our sample size was relatively small. This limited the statistical power and precision of our estimated effects. Second, we observed a lower-than-expected proportion of foals treated with macrolides alone or in combination with rifampicin, compared with findings previously reported12,23 and our personal experience with > 40% of foals commonly treated with antimicrobials at farms that use TUS for early detection of R equi pneumonia. To exclude underreporting as an explanation, investigators (NMS and MEG) verified the numbers of treated foals at several farms by reviewing records and interviewing attending veterinarians and farm personnel. Thus, we believe that the data collected in the present study were accurate and likely reflected a reduced frequency of treating foals secondary to an increased awareness of the risk of antimicrobial resistance emergence. Third, we did not verify the macrolide and rifampin treatment of foals. Thus, it was possible that the estimated numbers of foals treated at the farms were inaccurately reported by the farms. Fourth, we did not specifically determine that each foal that underwent TUS was subsequently treated with a macrolide, alone or in combination with rifampin; however, we demonstrated that the proportion of foals treated with antimicrobials was significantly higher at TUS farms than at non-TUS farms. Moreover, 2 of the authors (MEG and NMS) indicated that a standard practice at TUS farms is to implement treatment when detected lesions exceed a minimum threshold in foals that are not showing abnormal clinical signs.
Results of the present study were important regarding antimicrobial resistance in equine practice. Our results indicated that farms that rely on TUS and that consequently treat larger numbers of foals with antimicrobials generally have higher concentrations of MRRE in the soil. We do not wish to derogate the value of TUS for purposes of diagnosing or monitoring abnormalities or as a component of a screening program that could be more specific than the use of TUS alone. Considered with findings of prior studies,12,13,14,15,28 the results of the present study indicated that it is of utmost importance to emphasize prudent use of antimicrobials at horse-breeding farms and to identify effective alternative treatments for subclinical pneumonia detected by TUS and attributed to R equi.29 Epidemiological studies are needed to determine the duration of MRRE persistence on farms where it is identified and whether this antimicrobial resistance is shared by other bacteria that might serve as a repository of antimicrobial-resistance genes that could threaten public health.
Acknowledgments
Supported by the Grayson-Jockey Club Research Foundation. Maggie Greiter was supported by the Hagyard Equine Medical Institute. Noah Cohen was supported by the Link Equine Research Endowment at Texas A&M University. Cody Dailey was supported by the National Science Foundation under grant No. DGE-1545433.
The authors declare that there were no conflicts of interest.
This work is dedicated to the memory of our beloved colleague, Dr. Steeve Giguère.
Abbreviations
| CFU | Colony-forming unit |
| MRRE | Macrolide- and rifampin-resistant Rhodococcus equi |
| NANAT | Nalidixic acid, novobiocin, actidione, and potassium tellurite |
| TUS | Thoracic ultrasonographic screening |
Footnotes
Stata Statistical Software, release 14, Stata Corp LLC, College Station, Tex.
ETEST strips, bioMérieux Inc, Durham, NC.
RStudio, version 1.2.5033, RStudio, Boston Mass. Available at: rstudio.com. Accessed Nov 30, 2019.
clusterPower: Power Calculations for Cluster-Randomized and Cluster-Randomized Crossover Trials, version 0.6.111. Available at: CRAN.R-project.org/web/packages/clusterPower/clusterPower.pdf. Accessed Nov 30, 2019.
Pinheiro J, Bates D, DebRoy S, et al. nlme: Linear and nonlinear mixed effects models, version 3.1-148, Available at: CRAN.R-project.org/package=nlme. Accessed Nov 30, 2019.
Pinheiro J, Bates D, DebRoy S, et al. multcomp: Simultaneous inference in general parametric models, version 1.4-13. Available at: CRAN.R-project.org/package=multcomp. Accessed Nov 30, 2019.
References
- 1. ↑
Donisi A, Suardi MG, Casari S, et al. Rhodococcus equi infection in HIV-infected patients. AIDS 1996;10:359–362.
- 2. ↑
Yamshchikov AV, Schuetz A, Lyon GM. Rhodococcus equi infection. Lancet Infect Dis 2010;10:350–359.
- 3. ↑
Giguère S, Cohen ND, Chaffin MK, et al. Rhodococcus equi: clinical manifestations, virulence, and immunity. J Vet Intern Med 2011;25:1221–1230.
- 4. ↑
Giguère S, Gaskin JM, Miller C, et al. Evaluation of a commercially available hyperimmune plasma product for prevention of naturally acquired pneumonia caused by Rhodococcus equi in foals. J Am Vet Med Assoc 2002;220:59–63.
- 5. ↑
Ainsworth DM, Eicker SW, Yeagar AE, et al. Associations between physical examination, laboratory, and radiographic findings and outcome and subsequent racing performance of foals with Rhodococcus equi infection: 115 cases (1984–1992). J Am Vet Med Assoc 1998;213:510–515.
- 6. ↑
Venner M, Kerth R, Klug E. Evaluation of tulathromycin in the treatment of pulmonary abscesses in foals. Vet J 2007;174:418–421.
- 7. ↑
Slovis NM, McCracken JL, Mundy G. How to use thoracic ultrasound to screen foals for Rhodococcus equi at affected farms, in Proceedings. 51st Annu Conv Am Assoc Equine Pract 2005;274–278.
- 8. ↑
Chaffin MK, Cohen ND, Bloggett GP, et al. Evaluation of ultrasonographic screening parameters for predicting subsequent onset of clinically apparent Rhodococcus equi pneumonia in foals, in Proceedings. 59th Annu Conv Am Assoc Equine Pract 2013;268–269.
- 9. ↑
Venner M, Rödiger A, Laemmer M, et al. Failure of antimicrobial therapy to accelerate spontaneous healing of subclinical pulmonary abscesses on a farm with endemic infections caused by Rhodococcus equi. Vet J 2012;192:293–298.
- 10. ↑
Giguère S, Prescott JF. Clinical manifestations, diagnosis, treatment, and prevention of Rhodococcus equi infections in foals. Vet Microbiol 1997;56:313–334.
- 11. ↑
Giguère S, Lee E, Williams E, et al. Determination of the prevalence of antimicrobial resistance to macrolide antimicrobials or rifampin in Rhodococcus equi isolates and treatment outcome in foals infected with antimicrobial-resistant isolates of R equi. J Am Vet Med Assoc 2010;237:74–81.
- 12. ↑
Burton AJ, Giguère S, Sturgill TL, et al. Macrolide- and rifampin-resistant Rhodococcus equi on a horse breeding farm, Kentucky, USA. Emerg Infect Dis 2013;19:282–285.
- 13. ↑
Huber L, Giguère S, Cohen ND, et al. Identification of macrolide- and rifampicin-resistant Rhodococcus equi in environmental samples from equine breeding farms in central Kentucky during 2018. Vet Microbiol 2019;232:74–78.
- 14. ↑
Huber L, Giguère S, Cohen ND, et al. Prevalence and risk factors associated with emergence of Rhodococcus equi resistance to macrolides and rifampicin in horse-breeding farms in Kentucky, USA. Vet Microbiol 2019;235:243–247.
- 15. ↑
Huber L, Giguère S, Slovis NM, et al. Emergence of resistance to macrolides and rifampicin in clinical isolates of Rhodococcus equi from foals in central Kentucky, USA: 1995 to 2017. Antimicrob Agents Chemother 2018;63:e01714–e01718.
- 16. ↑
Ladrón N, Fernández M, Agüero J, et al. Rapid identification of Rhodococcus equi by a PCR assay targeting the choE gene. J Clin Microbiol 2003;41:3241–3245.
- 17. ↑
Giguère S, Hondalus MK, Yager JA, et al. Role of the 85-kilobase plasmid and plasmid-encoded virulence-associated protein A in intracellular survival and virulence of Rhodococcus equi. Infect Immun 1999;67:3548–3557.
- 18. ↑
Clinical and Laboratory Standards Institute. Performance standards for antimicrobial disk and dilution susceptibility tests for bacteria isolated from animals; approved standards. CLSI document VET01–A4. Wayne, Pa: CLSI, 2013.
- 19. ↑
Riesenberg A, Feßler AT, Frömke C, et al. Harmonization of antimicrobial susceptibility testing by broth microdilution for Rhodococcus equi of animal origin. J Antimicrob Chemother 2013;68:2173–2175.
- 20. ↑
Berghaus LJ, Giguère S, Guldbech K, et al. Comparison of Etest, disk diffusion, and broth macrodilution for in vitro susceptibility testing of Rhodococcus equi. J Clin Microbiol 2015;53:314–318.
- 21. ↑
Papich MG. In: Riviere JE, Papich MG, eds. Veterinary pharmacology and therapeutics. 10th ed. Wiley-Blackwell, 2017;903–952.
- 22. ↑
Venner M, Reinhold B, Beyerbach M, et al. Efficacy of azithromycin in preventing pulmonary abscesses in foals. Vet J 2009;179:301–303.
- 23. ↑
Chaffin MK, Cohen ND, Martens RJ, et al. Evaluation of the efficacy of gallium maltolate for chemoprophylaxis against pneumonia caused by Rhodococcus equi infection in foals. Am J Vet Res 2011;72:945–957.
- 24. ↑
Venner M, Astheimer K, Lämmer M, et al. Efficacy of mass antimicrobial treatment of foals with subclinical pulmonary abscesses associated with Rhodococcus equi. J Vet Intern Med 2013;27:171–176.
- 25. ↑
Willingham-Lane JM, Berghaus LJ, Berghaus RD, et al. Effect of macrolide and rifampin resistance on fitness of Rhodococcus equi. Appl Environ Microbiol 2019;85:e02665–e18.
- 26. ↑
Willingham-Lane JM, Berghaus LJ, Berghaus RD, et al. Effect of macrolide and rifampin resistance on fitness of Rhodococcus equi during intramacrophage replication and in vivo. Infect Immun 2019;87:e00281–e19.
- 27. ↑
Levin BR, Perrot V, Walker N. Compensatory mutations, antibiotic resistance and the population genetics of adaptive evolution in bacteria. Genetics 2000;154:985–997.
- 28. ↑
Anastasi E, Giguère S, Berghaus LJ, et al. Novel transferable erm(46) determinant responsible for emerging macrolide resistance in Rhodococcus equi. J Antimicrob Chemother 2016;71:1746.
- 29. ↑
Cohen ND, Slovis NM, Giguère S, et al. Gallium maltolate as an alternative to macrolides for treatment of presumed Rhodococcus equi pneumonia in foals. J Vet Intern Med 2015;29:932–939.
