Thomas R. Shryock, Donald W. White, J. Mitchell Staples, Carolyn S. Werner

J Vet Diagn Invest 8:337-344 (1 996) Minimum inhibitory concentration breakpoints and disk diffusion inhibitory zone interpretive criteria for tilmicosin susceptibility testing against Pasteurella spp. associated with bovine respiratory disease Thomas R. Shryock, Donald W. White, J. Mitchell Staples, Carolyn S. Werner Abstract. Tilmicosin is a novel macrolide antibiotic developed for exclusive use in veterinary medicine. The first tilmicosin-containing product was approved to treat bovine respiratory disease associated with pasteurellae. The development of antimicrobial susceptibility testing guidelines for tilmicosin was predicated on the relationship of clinical efficacy studies that demonstrated a favorable therapeutic outcome, on pharmacokinetic data, and on in vitro test data, as recommended by the National Committee for Clinical Laboratory Standards (NCCLS). The NCCLS-approved breakpoints for the MIC dilution testing are resistant 32 g/ml, intermediate 1 6 g/ml, and susceptible 8 g/ml. The zone of inhibition interpretive criteria for disk diffusion testing with a 1 5-g disk are resistant 1 0 mm, intermediate 1 1-1 3 mm, and susceptible 1 4 mm. Tilmicosin (20-deoxo-20-(3,5-dimethylpiperidin-1 - y1 )desmycosin) is a semisynthetic macrolide antibiotic 4,1 3,23 developed exclusively for use in livestock. The spectrum of activity of tilmicosin is predominantly against Gram-positive bacteria, with activity against the Gram-negative bacteria limited to members of the HAP group (Haemophilus, Actinobacillus, and Pasteurella) and mycoplasma. 23 The potential application of tilmicosin for the treatment of bovine pneumonia associated with pasteurellae was confirmed in 2 infection model studies. One experiment using a calf infection model with Mycoplasma bovis and Pasteurella haemolytica serotype 1 indicated that a single 2 0 mg/kg dose of tilmicosin was effective in decreasing clinical scores and pneumonic consolidation. 11 The second experiment was a natural infection model of calves with bovine respiratory disease (B RD), where a single administration of 1 0 mg/kg of tilmicosin was effective in lowering mortality and temperature, improving clinical signs, and reducing bacterial recovery from the lungs. 24 A preliminary report on the pharmacokinetics of tilmicosin in calves suggested that the whole-lung homogenate concentration of tilmicosin exceeded the minimum inhibitory concentration (MIC) for Pasteurella haemolytica for 72-96 hours posttreatment. A long-acting single subcutaneous administration therapy for treatment of B RD would be of value to the cattle industry by decreasing the number of injections needed for treatment, reducing handling and movement of sick animals, and avoiding the possibility of injection site lesions in edible tissue. Consequently, a series of worldwide clinical field studies demonstrated the beneficial therapeutic outcome with tilmicosin for the treatment of BRD. 6,10,14,15,17,18 31 The prudent use of tilmicosin a for the treatment of B RD is dependent on a collaborative relationship between the veterinary diagnostic laboratory, the attending veterinarian, and the producer. To assist the veterinarian, the laboratory must perform antimicrobial susceptibility testing properly and provide consistent interpretations of the results. Guidelines for the development of the interpretations of in vitro susceptibility test results for human pathogens have been described by the National Committee for Clinical Laboratory Standards (NCCLS), 20 although a separate subcommittee is now formulating specific veterinary guidelines. 19,33 The objective of the present study was to established MIC breakpoints and disk diffusion interpretive criteria for tilmicosin to enable the diagnostic laboratory to fulfill its role in the treatment process by providing meaningful information to the veterinary practitioner. Materials and methods Microorganisms. Clinical isolates (380) of Pasteurella haemolytica (n = 242) and P. multocida (n = 1 38) collected From Animal Science Discovery and Development, Elanco Animal Health, A Division of Eli Lilly and Co., Greenfield, IN 461 40 United States (TX, ID, OH, MO, OK) comprised the pop- within the last 1 0 yr from various geographic regions in the (Shryock, White, Staples), and CAVL, Inc., Amarillo, TX 791 1 8 ulation of organisms for evaluation. A minimum of 1 00-300 (Werner). isolates is recommended by the NCCLS to develop the interpretive criteria for a single genus or species. No special Received for publication July 31, 1 995. 337

338 Shryock et al. attempts were made to select for specific antimicrobial resistance patterns or mechanisms. Although Haemophilus somnus (another bovine respiratory pathogen) has tilmicosin MIC 90 s less than or equal to those of the pasteurellae, 33 the test methodology is sufficiently different (because of the fastidious nature of its growth requirements) that data for that microorganism were not included in the development of these interpretive criteria. Appropriate reference strains for quality control were used throughout the testing: Staphylococcus aureus ATCC 29213 (dilution tests), S. aureus ATCC 25923 (disk diffusion tests), and P. haemolytica 1 28K (a Lilly Research Laboratories reference strain). Antimicrobial susceptibility test methodology. The tilmieosin and erythromycin MICs were determined according to the procedures detailed by the NCCLS. 21 The broth microdilution technique was performed in custom made 96-well plates b using cation-adjusted Mueller-Hinton broth. c The agar dilution method employed 5% defibrinated sheep blood in Mueller-Hinton agar with antibiotic incorporated. Disk diffusion tests were conducted following the NCCLS recommendations. 22 Disks contained 1 5 g tilmicosin d and were placed on sheep blood agar (as above). Quality control guidelines for tilmicosin susceptibility testing have been described elsewhere; 27 however, the S. aureus ATCC 2 92 1 3 MIC guidelines have been modified by the NCCLS subcommittee to 1-4 g/ml. For comparison of the agar dilution and microbroth dilution assays of bovine pasteurellae against tilmieosin, 202 of the most recent isolates were selected. ph effect. The effect of ph on the antibacterial activity of tilmicosin was examined by determining the MIC for 29 isolates of P. haemolytica and P. multocida in Mueller-Hinton medium with the ph adjusted to 6.0, 7.0, or 8.0 with NaOH or HCl. Each medium was used to determine the MIC of the isolates using microbroth dilution techniques. Data analysis. The analysis of the scattergram data for determining the relationship between MICs and zone size was performed using formulae and guidelines previously described. 2,20,34 Results A scattergram to illustrate the relationship between MIC and zone size for a population of 380 P. haemolyticaand P. multocida isolates of bovine origin was constructed (Fig. 1 ). The isolates used in this study were best characterized as a randomly selected collection. MIC values were similar to those reported by others 12,32 Figure 1 shows a wide range of values, with a clustering of points near the center and scattered points away from the center. The uneven distribution of points over the range of concentrations tested was such that regression analysis was not possible, making error rate bounding the preferred statistic. 20,34 The outcome of the error rate bounding calculations is provided in Table 1 for the interpretive criteria cutoffs. Based on the total number of isolates tested, the P rate (population error rate) for very major errors (VME; those errors where an isolate is resistant by dilution testing but susceptible by disk testing, i.e., false-sus-

MIC breakpoints for Pasteurella spp. 339 ceptible disk test) was zero, for major errors (ME; those errors where an isolate is false resistant in the disk test) was zero, and for minor errors (those isolates classified as intermediate by 1 test but not the other) was 4.2%. The RC rate (risk corrected error rate) is based on only those isolates classified as resistant by the dilution method. Only 8 isolates were available for use in these calculations, rather than the 3 80 for the total population, and thus the RC rate statistic was not calculated. The effect of ph on the tilmicosin MIC result was investigated because the activity of macrolides is known to be influenced by ph. An 8-16 fold increase in the MIC (i.e., from 0.5 to 4 or 8 g/ml) was observed as the ph decreased from 8.0 to 7.0, with no growth observed at ph 6.0 (data not shown) for the pasteurellae. A comparison of the agar dilution method with the broth microdilution method was done to establish equivalency of results, as recommended by the NCCLS. 20 The MIC results are presented for each species of Pasteurella for frequency of occurrences at a given dilution end point (Table 2). Results for all but 6 of the 202 isolates were within the range of 2-8 g/ ml, with the majority at 2 or 4 g/ml. An MIC ratio calculation used previously 8 indicated that 87% of the paired isolate values (97% for P. multocida, 82% for P. haemolytica) had a ratio of 0.5, 1, or 2 indicating comparability, well within the accepted 1 dilution variation of test methodology. Tilmicosin and erythromycin are both members of the macrolide class of antibiotics. As specified in the NCCLS guidelines, 20 comparative studies within an antibiotic class are necessary. Accordingly, the 380 Pasteurella isolates were also tested by broth microdilution against erythromycin, with the results compared with tilmicosin MICs in a scattergram (Fig. 2). A regression analysis was not possible because of the clustered distribution of MIC values. Consequently, isolates were compared based upon the linear isolate pairing MIC ratio 8 of erythromycin : tilmicosin. Comparable MICs were observed for 98% of the isolates (excluding off scale or isolates with no end point). Figure 3 provides a comparison of the zones of erythromycin with those of tilmicosin. The observed trend was for the erythromycin zones to be 1-4 mm larger than the corresponding tilmicosin zones. Discussion The development of breakpoints and zone interpretive criteria is based on 3 components: therapeutic outcome of treatment, pharmacokinetics, and in vitro susceptibility testing. 20 For veterinary applications, the NCCLS M23-T2 can be used as a guide, but specific

340 Shryock et al. allowances for unique differences in livestock evaluations, as opposed to human patient populations, must be made. 19,33 The clinical evaluation was made of tilmicosin administered as a single 1 0 mg/kg subcutaneous treatment and compared to multiple intramuscular doses of other antibiotics to determine an effective therapeutic outcome. In a previous study, 31 calves with BRD that were treated once with tilmicosin had a significantly better response than those treated with ceftiofur as indicated by improved health, lower mortality, lower illness scores, and body temperature reduction. Eleven field studies in Canada evaluated tilmicosin and other antibiotics in feedlot BRD treatment. 17 A single dose of 1 0 mg/kg of tilmicosin was more effective than ceftiofur, oxytetracycline, or trimethoprim/sulfadoxine in reducing 1 ) mortality due to fibrinous pneumonia, 2) number of treatment days, and 3) re-pulls. A series of 8 field trials with naturally occurring BRD, conducted in the United S tates and Canada, demonstrated a significant reduction in mortality, temperatures, and clinical signs of disease and improved weight gains subsequent to tilmicosin treatment relative to placebo-treated controls. 10 Tilmicosin given on arrival at feedlots delayed the mean time to initial pneumonia treatment, allowed more weight gain, and significantly decreased mortality and morbidity attributable to BRD. 7,1 8 Isolates recovered from the lungs of dead animals did not exhibit resistance to tilmicosin. Clinical studies conducted outside of North America have shown similar favorable therapeutic outcomes. In tilmicosin trials in Europe, a dose of 1 0 mg/kg significantly reduced mortality and clinical illness in treated animals. 14 In a Hungarian study, a single administration of tilmicosin at 1 0 mg/kg was more effective against calf pneumonia than were other conventional therapies in decreasing mortality, reducing clinical signs, and restoring appetite in nearly 200 treated animals. 6 A successful treatment with 1 0 mg/kg of tilmicosin of a natural outbreak of pneumonia in calves in B ritain has

MIC breakpoints for Pasteurella spp. 341 also been described. 15 Together, these studies demonstrate that a single subcutaneous dose of 10 mg/kg tilmicosin a results in a beneficial therapeutic outcome in the treatment of BRD, thus establishing the first part of the necessary foundation for interpretive criteria development. The pharmacokinetics of tilmicosin in cattle have been studied using the single 1 0 mg/kg subcutaneous dose found to be efficacious in therapy. Data on tilmicosin in healthy cattle 30 showed that the serum tilmicosin concentrations never exceeded a mean level of 0.35 g/ml, whereas the mean lung tilmicosin concentration peaked at 7.1 7 g/g by 24 hours postadministration. S erum binding of tilmicosin is not substantial. 23 A more recent report measured the tilmieosin content of lung tissue obtained at various times posttreatment from calves with acute clinical pneumonia. 29 Mean lung homogenate tilmicosin concentrations increased to 9.30 g/g by 8 hours in normal lung tissue from steers with acute pneumonia, peaking at 9.78 g/g by 24 hours. A second study compared the tilmicosin content in different areas of the lung from the same anima1 28 (these animals also had severe clinical pneumonia). The nonconsolidated lung mean tilmicosin content reached 1 7.87 g/g by 8 hours posttreatment, whereas the consolidated portion of the lung tissue had a mean tilmicosin content of 9.50 g/g. Thus, tilmicosin appears to accumulate in normal lung tissue and is able to penetrate even consolidated tissue to high, effective levels. Alveolar macrophages, which are the resident phagocytic cells in the lung, accumulated tilmicosin to an extraordinary level (651 g/ml by 24 hours when incubated in medium containing 1 0 g/ml tilmicosin). This exceedingly concentrated amount of antibiotic could augment the ability of the phagocyte to destroy ingested bacteria within the phagolysosome. S everal human-use antibiotics indicated for treatment of respiratory infections have received FDA approval in part because of the high intraphagocytic cell concentration of the drug. 3 Intracellular transport of tilmicosin by leukocytes to specific infection sites within lung tissue and subsequent extracellular deliveryby efflux mechanisms may be similar to that proposed for azithromycin. 9 Thus, the local concentration of tilmicosin at the site of the bacterial infection could well exceed the concentrations measured in lung tissue homogenates. The data available on the pharmacokinetics of tilmicosin in the bovine lung provide the second part of the information needed to establish interpretive criteria. The third component for the proposed guidelines is antimicrobial susceptibility testing. The initial approach to scattergram evaluation is to use a population of isolates, regardless of the specific site of bodily origin, as described in the NCCLS M23-T2. 20 Thus, the population of isolates obtained for this study was comprised of clinical laboratory submissions from cattle with B RD (nasal and lung specimens). Comparison of the proposed breakpoints using pathogens isolated from animals in clinical trials would validate the test criteria and can be done at a later time, owing to the intricacies of such studies. For example, an evaluation of an appropriate isolate from a specific individual animal is not always practical in the clinical evaluation of B RD 33 mainly because of the difficulties in sample acquisition (e.g., performing a transtracheal washout in a calf in a feedlot situation). Even though nasopharyngeal cultures do not accurately predict the results of a bronchoalveolar culture 1 and one cannot be certain that the etiologic agent recovered from a nasal swab is in fact the same agent responsible for the pneumonia, these types of specimens are routinely collected pretreatment. Furthermore, in vitro antibiotic susceptibility testing (by disk diffusion) of P. haemolytica isolated from pretreatment nasal swabs was not a useful predictor of treatment success of the antibiotic. 16 While this limited usefulness was attributed in part to the lack of interpretive criteria available for the veterinary application of the antibiotic, it may also indicate the differences in isolate origin. 1 Obtaining lung samples from treated animals is also impractical and may not be truly representative of the clinical situation. Animals that recover can only be sampled by nasal swabs, whereas animals that die have theoretically already been exposed to the antibiotic, making a susceptibility test biased. Integrating the available data on therapeutic outcome, pharmacokinetics, and antimicrobial susceptibility testing provided a rational basis for the MIC breakpoints and zone interpretive criteria. A single subcutaneous dose of 1 0 mg/kg of tilmicosin resulted in an acceptable therapeutic outcome. The pharmacokinetic data for that dose demonstrated a peak wholelung homogenate mean tilmicosin concentration of 7.1 7 g/g at 2 4 hours in healthy calves and 9.50 g/g in consolidated lung tissue taken from calves with acute pneumonia. The MIC frequency results (Fig. 1 ) showed that 94% of the P. haemolytica and P. multocida isolates had end points 8 g/ml and therefore would be predicted to be fully susceptible to tilmicosin in vivo because the amount of tilmicosin available in either normal or consolidated lung tissue was near to or exceeded the MIC. Thus, isolates with MIC 8 g/ml can be considered susceptible to tilmicosin. A value of 16 g/ml as an intermediate classification is a conservative characterization. By definition, intermediate isolates can be considered to respond to therapy if the antibiotic is known to concentrate in a target tissue. 21 The 1 6 g/ml level is above the amount of tilmicosin achievable in normal, healthy lung tissue, below the

342 Shryock et al. 17.87 g/g peak level in nonconsolidated lung tissue and included tilmicosin. 32 The 4-year overall MIC 90 s and alveolar macrophages, but just over the peak con- for P. haemolytica, P. multocida, and Haemophilus centration observed for consolidated lung tissue. Sub- somnus were 8, 8, and 4 g/ml, respectively. Although MIC levels of tilmicosin have been shown to exert a mention is made of the original breakpoint scheme for postantibiotic effect on the pasteurellae, which could tilmicosin, the authors proposed their own breakpoints result in a weakened population of bacteria that are for tilmicosin to classify isolates as either susceptible more readily eliminated by host defenses. 5 Because the (MIC 4 g/ml) or resistant (MIC 8 g/ml), with ph of pleural fluid is generally around 7.0 or is slightly no provision made for an intermediate category. Howlower in a disease state (G. Brumbaugh, personal com- ever, in view of the breakpoints in this report (Fig. 1 ), munication), the MIC values generated in vitro should isolates with an MIC of 1 6 g/ml (intermediate) should be representative of the in vivo ph environment. A be considered as likely to respond to therapy; and inlower ph tends to abrogate activity. Thus, isolates ob- eluded with isolates with an MIC 8 g/ml (susceptained from an animal with B RD and classified as tible) in any evaluation of overall susceptibility. Only intermediate should still be expected to respond to those isolates with an MIC 32 g/ml should be contherapy because of the potential for the tissue and al- sidered resistant and refractory to therapy. Conseveolar macrophage concentrations of tilmicosin to ex- quently, interpretations of the percentage of tilmicosinteed 1 6 g/ml at local sites of infection. Thus, true susceptible isolates 32 underestimated the number of resistance assessments should only be based on those isolates that could have been considered capable of isolates with an MIC 3 2 g/ml because those isolates responding to tilmicosin treatment. Those MIC 90 s, in the intermediate category are predicted to be re- interpreted in the light of the approved breakpoints, sponsive to therapy in most cases. Isolates with an MIC indicate that 90% of the BRD isolates would be re- 32 g/ml are considered resistant because whole-lung sponsive to tilmicosin therapy, and only a small perhomogenate concentrations of tilmicosin are not reli- centage ( 1 0%) are truly resistant. The data further ably achieved using the currently approved dosage and indicate that this is the case for all 4 years of the survey thus the animal would not be expected to respond to (i.e., tilmicosin MIC 90 s were 8 g/ml) for each of the therapy. 3 BRD organisms, with the exception of an MIC 90 of Corresponding zone interpretive criteria for disk dif- 16 g/ml for P. multocida isolates in 1 990-1 991. Befusion testing was developed in association with the cause tilmicosin was not approved in the USA for use MIC tests. Preliminary data (not shown) indicated that in cattle until March 1 992, none of the US isolates a 1 5-g disk was acceptable. The ideal minimum rec- collected and tested in this survey year could ever have ommended zone size for susceptibility is 1 5-25 mm. 20 been exposed to the antibiotic. A comparison of til- However, the natural population distribution did not micosin MICs for US and Canadian isolates was conallow compliance with this concept; zones of inhibition ducted and interpreted 32 as suggesting that a larger 1 4 mm were considered susceptible when matched proportion of Canadian isolates than US isolates was against the respective MIC values. The zone size and susceptible to tilmicosin. Again, the interpretation of breakpoint cutoffs should be adjusted using the error isolates with an MIC 4 g/ml (susceptible), to derive rate bounded calculations so that VME and ME are the percentage susceptible, underestimated the number minimized, while maintaining relevance to the phar- of susceptible isolates. Based on the MIC 90 data promacokinetic data. For the P rate VME were zero, which vided, 1 00% of the US BRD isolates would be remeets the ideal value of 1 5%, ME were zero com- sponsive to tilmicosin therapy (susceptible or interpared with an allowable 3% rate, and the minor error mediate), with none truly resistant (i.e., MIC 3 2 g/ rate was 4.2%. The paucity of resistant strains available ml), a situation similar to that of the Canadian isolates to calculate the RC rate made this value questionable in (i.e., all of the isolates were susceptible); thus, no major its utility; thus, it was not used. As noted in the NCCLS differences between the percentages of nonresistant iso- M23-T2, 20 greater error rates were likely to result from lates from the 2 countries were evident. the use of a restricted population of isolates in this A comparison of the broth microdilution and agar calculation. Thus, the use of the P rate is a better indi- dilution procedures showed that both methods gave cator of the likelihood of errors than the RC rate in this comparable results, i.e., within 1 dilution of each other instance. Overall, the use of the error rate bounded for the majority of the isolates tested. Additionally, the method (as opposed to regression analysis) to evaluate use of the broth microdilution-based S ensititre system the relationship between the disk diffusion method and has been evaluated against an agar dilution technique the broth dilution method demonstrated the accepta- in a prior study and yielded comparable results. 25 Thus, bility of the criteria. an alternative method for conducting the in vitro test- A large survey of antimicrobial resistance trends for ing is available. isolates from cattle with B RD was recently completed S tudies comparing tilmicosin and erythromycin were

MIC breakpoints for Pasteurella spp. 343 performed. The MIC values of erythromycin and tilmicosin were very similar; however, breakpoint interpretive criteria for the 2 macrolides are vastly different. Erythromycin resistant breakpoints range from 8 g/ ml 21 to 4 g/ml 26 to 2 g/ml 32 compared with 32 g/ml for tilmicosin. Erythromycin susceptible breakpoints range from 0.5 g/ml 21 to 1.0 g/ml 32 compared with 8 g/ml for tilmicosin. Zone diameter interpretive criteria also are markedly different for the 2 macrolides, in spite of a similarity of zone inhibition diameters. The breakpoint and zone interpretive criteria for tilmicosin were based on therapeutic outcome data, pharmacokinetic studies, and in vitro susceptibility testing. This information will allow the veterinary diagnostic laboratory to provide information on susceptibility tests to veterinarians and improve the rational use of tilmicosin in the treatment of BRD. Acknowledgements We acknowledge the helpful comments of L. Stobbs, T. Thomson, D. Preston, and B. Martin of Lilly Research Laboratories and the expert administrative assistance of Ms. P. Lawhorn in the preparation of the manuscript. Sources and manufacturers Micotilfi, Elanco Animal Health, Indianapolis, IN. Sensititre, Westlake, OH. Becton-Dickinson Microbiology Systems, Cockeysville, MD. Difco, Detroit, MI. References 1. Allen, JW, Viel L, Bateman KG, et al.: 1 991, The microbial flora of the respiratory tract in feedlot calves: associations between naso-phargngeal and bronchoalveolar lavage culture. Can J Vet Res 55:341-346. 2. Bradford PA, Sanders CC: 1 992, Use of a predictor panel for development of a new disk for diffusion tests with cefoperazonesulbactam. Antimicrob Agents Chemother 36:394-400. 3. Butts JD: 1 994, Intracellular concentrations of antibacterial agents and related clinical impressions. Clin Pharmacokinet Concepts 2 7:63-84. 4. Debono M, Willard KE, Kirst HA, et al.: 1 989, Synthesis and antimicrobial evaluation of 20-deoxo-20-(3,5-dimethylpiperidin-1 -yl)desmycosin (tilmicosin, EL-870) and related cyclic amino derivatives. J Antibiot 1 3: 1 253-1 267. 5. Diarra MS, Malouin F: 1 994, Postantibiotic effects of tilmicosin and of other antibacterial agents at MICs and sub-mics on swine and bovine pathogens. Proc Annu Meet Am Soc Microbiol 94:1 6. [Abstr. A-81 ] 6. Fodor L, Varga J, Gallowitsch F, et al.: 1993, Treatment of calf pneumonia with tilmicosin. Acta Vet Hung 41 :41-49. 7. Galyean ML, Gunter SA, Malcolm-Callis KJ: 1 995, Effects of arrival medication with tilmicosin phosphate on health and performance of newly received beef-cattle. J Anim Sci 73:1 21 9-1226. 8. Gavan TL, Jones RN, Barry AL: 1 980, Evaluation of the Sensititre system for quantitative antimicrobial drug susceptibility testing: a collaborative study. Antimicrob Agents Chemother 1 7:464469. 9. Gladue RP, Brigh GM, Isaacson RE, Newborg MF: 1 989, In vitro and in vivo uptake of azithromycin (CP-62,993) by phagocytic cells: possible mechanism of delivery and release at sites of infection. Antimicrob Agents Chemother 33:277-282. 10. Gorham PE, Carroll LH, McAskill JW, et al.: 1 990, Tilmicosin as a single injection treatment for respiratory disease of feedlot cattle. Can Vet J 31 :82 6-82 9. 11. Gourlay RN, Thomas LH, Wyld SG, Smith C: 1 989, Effect of a new macrolide antibiotic (tilmicosin) on pneumonia experimentally induced in calves by Mycoplasma bovis and Pasteurella haemolytica. Res Vet Sci 47:84-89. 12 Hartman EG, Geryl J: 1 993, Comparison between the minimal inhibitory concentration of tilmicosin and oxytetracycline for bovine pneumonic Pasteurella haemolytica isolates. Vet Q 1 5: 184. 13. Kirst HA, Willard KE, Debono M, et al.: 1 989, Structureactivity studies of 20-deoxo-20-amino derivatives of tylosinrelated macrolides. J Antibiot 1 7:1 673-1 683. 14. Laven R, Andrews AH: 1 991, Long-acting antibiotic formulations in the treatment of calf pneumonia: a comparative study of tilmicosin and oxytetracycline. Vet Rec 129:109-111. 15. Lawrence K: 1 994, An evaluation of the efficacy of tilmicosin in the treatment of bacterial pneumonia in European cattle. Proc Int Congr Eur Assoc Vet Pharmacol Ther 6:000-000. 16. Mechor GD, Jim GK, Janzen EG: 1 988, Comparison of penicillin, oxytetracycline, and trimethoprim-sulfadoxine in the treatment of acute undifferentiated bovine respiratory disease. Can Vet J 29:438-443. 17. Merrill JK, Tonkinson LV: 1 989, The effectiveness of Micotil for the treatment of bovine respiratory disease. Bovine Pract 24:26-28. 18. Merck DW, Merrill JK, Thorlakson BE: 1 993, Prophylactic efficacy of tilmicosin for bovine respiratory tract disease. J Am Vet Med Assoc 202:273-277. 19. Mortensen JE, Shryock TR: 1 994, Veterinary antimicrobial susceptibility testing coming of age. Clin Microbiol Newsl 1 6( 1 6): 134-135. 20. National Committee for Clinical Laboratory Standards: 1992, Development of in vitro susceptibility testing criteria and quality control parameters, 2 nd ed., Tentative guideline M2 3-T2. NCCLS, Villanova, PA. 21. National Committee for Clinical Laboratory Standards: 1993, Methods for dilution antimicrobial susceptibility tests for bacteria that grow aerobically, 3rd ed., Approved standard M7-A3. NCCLS, Villanova, PA. 22. National Committee for Clinical Laboratory Standards: 1993, Performance standards for antimicrobial disk susceptibility tests, 5th ed., Approved standard M2-A5. NCCLS, Villanova, PA. 23. Ose EE: 1 987, In vitro antibacterial properties of EL-870, a new semi-synthetic macrolide antibiotic. J. Antibiot 1 5:1 90-194. 24. Ose EE, Tonkinson LV: 1 988, Single-dose treatment of neonatal calf pneumonia with the new macrolide antibiotic tilmieosin. Vet Rec 123:367-369. 25. Phillips I, Warren C, Waterworth P: 1 978, Determination of antibiotic sensitivities by the Sensititre system. J Clin Pathol 31:531-535. 26. 27. 28. Post KW, Cole NA, Raleigh RH: 1 991, In vitro antimicrobial susceptibility of Pasteurella haemolytica and Pasteurella multocida recovered from cattle with bovine respiratory disease complex. J Vet Diagn Invest 3: 124-126. Shryock TR, White DW, Werner CS, Staples JM: 1 995, Proposed quality control guidelines for antimicrobial susceptibility tests using tilmicosin. J Clin Microbiol 33:331-335. Thomson TD, Laudert SB, Chamberland S, Lawrence K: 1994, Micotil-pharmacokinetics of tilmicosin, a semi-synthetic mac-

3 44 Shryock et al. 29. 30. 31. 32. rolide antibiotic. Proc Int Congr Eur Assoc Vet Pharmacol. Ther 6. Thomson TD, Lawrence K: 1 994, Micotil: pharmacokinetics of tilmicosin, a semi-synthetic macrolide, in cattle. Int Congr 33. Eur Assoc Vet Pharmacol Ther 6. Thomson TD, Matsuoka T: 1 989, Correlation of lung concentrations of a new veterinary antimicrobial with efficacy against Pasteurella haemolytica pneumonia in feedlot cattle. Proc Annu 34. Meet Conf Res Workers Anim Dis 70:1 99. Van Koevering MT, Smith RA, Gill DR, Owens FN: 1 992, Effects of tilmicosin or ceftiofur sodium on health and performance of shipping-stressed calves. J Anim Sci 70:90. Watts JL, Yancy RJ Jr, Salmon SA, Case CA: 1 994, A 4-year survey of antimicrobial susceptibility trends for isolates from cattle with bovine respiratory disease in North America. J Clin Microbiol 32:725-731. Watts JT, Yancy RJ Jr: 1 994, Identification of veterinary pathogens by use of commercial identification systems and new trends in antimicrobial susceptibility testing of veterinary pathogens. Clin Microbiol Rev 7:346-356. Woolfrey BF, Fox JM, Quall CO, Lally RT: 1 983, Error rates associated with the use of recently proposed breakpoints for testing Pseudomonas aeruginosa versus gentamicin, tobramytin, and amikacin by the standardized disk agar diffusion test. Antimicrob Agents Chemother 24:764-770.