Use of Susceptibility Testing in Veterinary Medicine

Peter D. Constable, BVSc, MS, PhD, DACVIM Department of Veterinary Clinical Medicine, University of Illinois, Urbana, IL 61 802

Abstract three categories, susceptible (sensitive), intermediate and resistant, using recommendations from the National There has been increased interest in optimizing Committee for Clinical Laboratory Standards (NCCLS) treatment protocols for antimicrobial agents, with sub­ for testing Veterinary Pathogens.4 The intermediate stantial reliance on susceptibility testing of bacterial term indicates an MIC value that is close to the pathogens isolated from diseased cattle. Antimicrobial breakpoint.5 susceptibility testing of bovine bacterial pathogens has The broth dilution method provides a direct mea­ traditionally used the agar diffusion (Kirby-Bauer) surement of MIC, and determines the ability of the method, which was designed to reflect the antibiotic pathogen to grow in the presence of a known antibiotic concentration in serum and interstitial fluid of human concentration. The broth dilution test is usually per­ patients. The validity of agar diffusion susceptibility formed as a commercially available microdilution test breakpoints derived from humans to the treatment of (Sensititre, Westlake, OH) in a 96 well microtiter plate mastitis, diarrhea and respiratory disease in cattle has that permits the testing of 12 antibiotics in a range of not been established. The use of susceptibility testing eight 2-fold dilutions.5 The microdilution method starts to guide treatment decisions for individual cattle is not by using a sterile loop to remove 3-5 representative colo­ recommended until the breakpoints have been validated nies from a 24-hour bacterial culture plate (use of mul­ as being predictive of treatment outcome. tiple colonies avoids selection of an atypical variant). The bacterial colonies are then suspended in 5 ml of .g (D Introduction 0.9% NaCl and the bacterial suspension standardized l::l ~ by adjusting the turbidity to a 0.5 McFarland standard (") (") 7 (D A number of methods have been used to determine (an index of bacterial concentration approximating 10 V, the susceptibility of bovine pathogens to antimicrobial colony forming units [CFU]/ml). A fixed volume aliquot V, agents: broth dilution, milk dilution (for mastitis patho­ is then transferred to Mueller-Hinton broth, and an gens), agar dilution, determination of mean bactericidal automated inoculation device used to deposit a standard­ concentration, determination of killing kinetics and the ized inoculum into each well of the microtitration tray agar diffusion method (Kirby-Bauer test). The first five containing a geometric progression of dehydrated anti­ methods are quantitative, whereas the agar diffusion microbial agent concentrations. Growth is recorded by method is qualitative.15 Because of issues related to cost monitoring the turbidity of each well, and the first dilu­ and complexity, the broth microdilution method is the tion with non-visible growth considered to be the MIC recommended gold standard method for in vitro suscep­ for that isolate.41 tibility testing, whereas agar diffusion provides a crude, The agar diffusion method is also called the Kirby­ inexpensive and clinically practical method for deter­ Bauer method, and the test procedure has changed little mining in vitro susceptibility. since Bauer, Kirby and others standardized the method Two important concepts (minimum inhibitory con­ in 1966.4•7 Three to 10 representative bacterial colonies centration [MIC], and breakpoints for MIC) need to be are selected from a blood agar plate and suspended in a understood when interpreting the results of suscepti­ fixed volume of sterile 0.9% NaCl to achieve the turbid­ bility testing. The minimum inhibitory concentration is ity of a 0.5 McFarlane standard; the suspension is then the lowest antibiotic concentration (expressed in µg/ml) spread evenly across the surface of an agar plate using that, under defined in vitro conditions, prevents the a sterile cotton swab. Small circular disks of filter pa­ growth of bacteria within a defined period oftime.5 It is per or tablets impregnated with antimicrobial agents generally accepted that MIC values are very repeatable. 6 are placed on the agar plate using flamed forceps or a

Statistics such as MIC50 (the median MIC for all iso­ special applicator and gently pressed down to ensure lates) and MIC90 (the MIC value which exceeds or equals contact.7 The agar plate is selected based on the bacte­ the MIC for 90% of the isolates) are frequently used to rial species being tested and incubated at 37°C over­ summarize population data. Breakpoints for MIC are night. During incubation, antibiotics dissolve from the specific MIC values used to assign bacteria to one of filter paper or tablets into the surrounding agar and

11 TH E AABP PROCEEDINGS-VOL. 37 thereby inhibit bacterial growth. The diameter of the eed publication documenting the relationship between zone of inhibition (Figure 1) is measured in mm and is the MIC of bacteria isolated from the site of infection correlated in some manner with the MIC for the bacte­ and clinical outcome in individual cattle administered (Q) ria.4 Interpretative zone diameters differ for each anti­ an antimicrobial agent. Accordingly, the recommended n 0 biotic because of differences in MIC and the diffusion MIC breakpoints appear to be based on in vitro MIC -0 '-< and solubility of the antibiotic in agar. Because agar values, pharmacokinetic/pharmacodynamic data and the "-1 cjq" diffusion is qualitative, the method is inferior to the results of clinical trials indicating efficacy using a stated ;::!"' quantitative broth dilution method. dosage protocol. In other words, we are still using the ...... approach described by Hjerpe in his seminal 1976 pa­ Determination and validation of per on the treatment of bacterial pneumonia in cattle.20 susceptibility breakpoints Many problems exist with the currently used sus­ ceptibility breakpoints for bacteria isolated from cattle. A standardized testing procedure for determining Accurate antimicrobial susceptibility test breakpoints antimicrobial susceptibility has been developed in the should be derived using MIC values for 300 to 600 iso­ United States by the NCCLS. A veterinary subcommit­ lates from representative clinical cases from a large geo­ tee ofNCCLS, called the Veterinary Antimicrobial Sus­ graphic area, 4·6 published pharmacokinetic/ ceptibility Testing (VAST) subcommittee, was formed in pharmacodynamic data for cattle, and clinical and bac­ 1992. The VAST subcommittee published a proposed teriologic cure rates.6 The results of field studies that standard in 1994 and approved standards in 1999 and measure the rate of clinical cure, using clinically rel­ 2002. 4 The VAST committee recommends an official in­ evant end points such as mortality, weight gain, treat­ terpretative MIC breakpoint against specific bacteria ment duration and relapse rate should be reported as a at a stated dosage protocol for a specific disease in a bare minimum. The rate of bacteriologic cure within a species. The MIC breakpoint is determined by consid­ specified time interval, using biologically relevant end­ ering available in vitro susceptibility data, pharmaco­ points such as failure to isolate the same pathogen from kinetic/pharmacodynamic data, and clinical efficacy the affected quarter in cows with mastitis, the feces in data; however, this author is unaware of a single refer- calf diarrhea, or a transtracheal wash in pneumonia,

Figure 1. Antimicrobial susceptibility testing using the agar diffusion (Kirby-Bauer) test. The diameter of the zone of inhibition around each antibiotic disk is associated with the MIC value for each antibiotic. Interpretative zone diameters differ for each antibiotic because of differences in the MIC value, diffusion rate and solubility of the antibiotic in agar. Photograph courtesy of Dr. DE Morin.

SEPTEMBER, 2004 12 also provides useful data. Clinical and bacteriologic cure trimethoprim/sulfadiazine.4 Susceptibility testing there­ rates may provide a clear breakpoint, or in other situa­ fore usually does not employ the active antibiotic agents tions, this data can be used in conjunction with phar­ present in commercially available antibiotic treatments (Q) macokinetic/pharmacodynamic data to suggest the most for cattle. It is likely that susceptibility test results based n 6 0 appropriate breakpoint. Unfortunately, the ideal ap­ on class representatives rather than the active antibi­ "O otic agent will lead to erroneous results. '< proach to determine accurate susceptibility breakpoints ....."'"I in cattle is hampered by three main difficulties: 1) lim­ {IQ g' ited availability of contemporaneous MIC values from Susceptibility testing in mastitis, calf diarrhea heterogeneous geographic locations, 2) incomplete and pneumonia ► ~ phamacokinetic/pharmacodynamic data, and 3) com­ ....."'"I (') plete absence of published field studies validating the Important considerations for treating bacterial § suggested susceptibility breakpoints. The effect of dis­ diseases in cattle are: 1) administering an antibiotic as Cfl► ease on the pharmacokinetics of antimicrobial agents directed on the label whenever possible, 2) using an Cfl 0 has usually been ignored, but differences in the plasma antimicrobial agent with an appropriate spectrum of (') 3 12 ~- concentration-time profile for oxytetracycline • and activity, 3) selecting an antimicrobial agent that attains ,.... erythromycin9 exist between healthy and pneumonic and maintains an effective therapeutic concentration at 0 :::1 calves, and experimental induction of endotoxemia and the site of infection, 4) treating for an appropriate dura­ 0 H; pyrexia in suckling Holstein calves changes the plasma tion, and 5) avoiding adverse local or systemic effects to 0 concentration-time profile for orally or intramuscularly and violative residues. Exisiting limitations in suscep­ < administered amoxycillin trihydrate.19 Finally, antimi­ tibility testing will be illustrated by examining three s· (1) crobial agents may produce beneficial effects separate economically important and common bacterial diseases to their activity against bacteria. A recent example of of cattle, namely mastitis, calf diarrhea and pneumo­ ~ ('),....,. this is the observation that tilmicosin decreases pulmo­ nia. ,....,.,.... nary inflammation in bovine pneumonia; this anti-in­ §i' (1) flammatory effect appears to be mediated, in part, by Mastitis "'"I Cfl inducing neutrophil apoptosis11 and by decreasing mac­ Antimicrobial agents are often selected based on 0 rophage phospholipase A2 activity and macrophage pro­ availability of labeled drugs, clinical signs in the cow, "O (1) 21 duction of PGE2• milk culture results for previous mastitis episodes in :::1 Specific antibiotics that are considered represen­ the herd, experience of treatment outcome in the herd, ~ (') (1) tative of their class are routinely used in susceptibility treatment cost and withdrawal times for milk and Cfl Cfl testing. For instance, cephapirin is a first generation slaughter.14•15•24 The validity of agar diffusion suscepti­ &. (narrow spectrum) cephalosporin used to treat mastitis bility breakpoints derived from humans to the treatment ,....,.Cfl in dairy cows. However, susceptibility testing does not of bovine mastitis has not been established and is ex­ ;.:"'"I use cephapirin as the test antibiotic, instead, cephal­ tremely questionable because bovine milk pH, electro­ a,.... othin is used because it is the recommended represen­ lyte, fat, protein, and leukocyte concentrations, growth 0p tative of first generation compounds.1 This factor composition, and pharmacokinetic profiles are recommendation ignores the results of an in vitro study different than those for human plasma, and because indicating one dilution difference in MIC50 values be­ human bacterial pathogens are often different from bo­ tween cephapirin and cephalothin for Staphylococcus vine mastitis pathogens. Also, antibiotics are distrib­ aureus, Proteus mirabilis and Citrobacter spp,40 and ig­ uted unevenly in an inflamed gland, and high antibiotic nores general recommendations that the antibiotic to concentrations can alter neutrophil morphology or func­ be used clinically should be tested.5 Another example is tion in vitro and thereby inhibit bacterial clearance in desfuroylceftiofur, the active metabolite of ceftiofur in vivo. 15 cattle. Desfuroylceftiofur and ceftiofur have similar We do not currently have adequate databases of in MICs for gram negative pathogens, whereas vitro MIC values for clinical mastitis pathogens, al­ desfuroylceftiofur has 2-3 serial dilutions less antimi­ though adequate databases are available for subclini­ crobial activity than ceftiofur against Staphylococcus cal mastitis isolates. Although we have good knowledge aureus, coagulase negative Staphylococcus spp, Strep­ regarding the pharmacokinetics of many parenteral tococcus uberis and Streptococcus dysgalactae.31 The antibiotics used to treat clinical mastitis, most pharma­ VAST subcommittee of the NCCLS recommended that cokinetic data has been obtained in healthy cattle and tetracycline be used as the class representative for ox­ it has not been determined whether pharmacokinetic ytetracycline, ampicillin be used to test for susceptibil­ values in healthy cows are the same as those in cows ity to amoxicillin and hetacillin, oxacillin be used to test with clinical mastitis. In addition, pharmacokinetic val­ for susceptibility to cloxacillin and trimethoprim/ ues for many of the intramammary antibiotics used to sulfamethoxazole be used to test for susceptibility to treat clinical mastitis are unknown, and we have a lim-

13 THE AABP PROCEEDINGS-VOL. 37 ited understanding of the pharmacodynamics of antibi­ Calf diarrhea otics in treating mastitis. More importantly, the The most important determinant of antimicrobial breakpoints currently recommended for all parenterally efficacy in treating calf diarrhea is obtaining an effec­ and almost all intramammary administered antibiotics tive antimicrobial concentration against bacteria at the are based on achievable serum and interstitial fluid con­ two sites of infection (small intestine and blood). The centrations in humans after oral or intravenous antibi­ results offecal antimicrobial susceptibility testing have otic administration. The relevance of these breakpoints traditionally been used to guide treatment decisions; to achievable milk concentrations in lactating dairy cows however, susceptibility testing in calf diarrhea probably after intramammary, subcutaneous, intramuscular, or has clinical relevance only when applied to fecal iso­ intravenous administration is dubious at best. We re­ lates of enterotoxigenic strains of E. coli or pathogenic cently demonstrated that the recommended breakpoints strains of Salmonella spp, and blood culture isolates for cephapirin and oxytetracycline were not predictive from calves with bacteremia. Validation of susceptibil­ of treatment outcome in cows with clinical mastitis.14 ity testing as being predictive of treatment outcome for In vitro antimicrobial susceptibility test results calves with diarrhea is currently lacking. have been suspected to be poorly correlated with treat­ Susceptibility testing in calf diarrhea has focused ment outcome for clinical mastitis and the value of the on using fecal isolates, although this approach is funda­ agar diffusion method to guide mastitis treatment deci­ mentally flawed. There do not appear to be any data sions has been widely questioned.15•26•30 Susceptibility demonstrating that fecal bacterial flora is representa­ testing in artificial media (agar or broth) cannot simu­ tive of small intestinal bacterial flora, which is the site late what happens in the mammary gland milk phase of infection in the intestinal tract of calves with of cows with mastitis and therefore cannot predict the enterotoxigenic E. coli. 13 Moreover, the predominant outcome oftherapy.26 With the exception of pirlimycin39 strain of E. coli in the feces of a scouring calf usually and penicillin-novobiocin,38 zone diameters in the agar changes during the diarrhea episode,34•35 and 45% (9/ diffusion test have not been related to antibiotic con­ 20) of diarrheic calves have different strains of E. coli centrations achieved in the bovine mammary gland with isolated from the upper and lower small intestine.34 In dosage regimens used by veterinarians and dairy pro­ other words, fecal E. coli strains should not be consid­ 0 ered to be representative of small intestinal E. coli "O ducers. It must be emphasized that the preliminary (D breakpoints for pirlimycin and penicillin-novobiocin strains. :::s combination have not been confirmed because of a lack A clear bias present in most antimicrobial suscep­ ~ () (D of appropriate field efficacy data with adequate num­ tibility studies conducted on fecal E. coli isolates is that V, V, 28 41 bers of susceptible and resistant isolates. • Moreover, data is usually obtained from dead calves, which are 0.. the breakpoints were calculated from phamacokinetic likely to be treatment failures. Calves that die from di­ v_;· =t data obtained in healthy mammary glands; it is likely arrhea are likely to have received multiple antimicro­ &= that pharmacokinetic values differ in glands with clini­ bial treatments, and preferential growth of antimicrobial s. cal mastitis. resistant E. coli strains starts within three hours of an­ 0 ?-· Results from field studies are available to evaluate timicrobial administration.27 Similar to susceptibility the validity of susceptibility breakpoints in guiding treat­ testing of mastitis pathogens, the agar diffusion break ment of cows with clinical or subclinical mastitis. The points for susceptibility testing of fecal isolates are not results from these field studies suggest that the follow­ based on achievable antimicrobial concentrations in the ing antibiotics may have valid (but not necessarily opti­ small intestine and blood of calves, but on achievable mal) breakpoints for treating clinical or subclinical antimicrobial concentrations in the plasma of humans. mastitis caused by specific bacteria; parenteral penicil­ The only study to statistically test the predictive lin G for subclinical Staphylococcus aureus infections, ability offecal antimicrobial susceptibility results found intramammary cephapirin for clinical Streptococcus spp that the rectal swab was an inaccurate method of pre­ infections, and parenteral trimethoprim-sulfadiazine for dicting clinical outcome.10 Two reports concluded that clinical Escherichia coli infections. Of these three antibi­ a "good correlation" existed between in vitro antimi­ otics, the breakpoints for penicillin G and cephapirin have crobial susceptibility of fecal E. coli isolates and clini­ only been validated for bacteriologic cure, whereas the cal response to antimicrobial treatment; however, breakpoint for trimethoprim-sulfadiazine is validated for neither study statistically tested the association.33•36 In clinical cure.15 Because duration of infection before treat­ contrast, two other studies reported "no correlation" ment, antibiotic dosage, dosage interval and duration of between in vitro antimicrobial susceptibility of fecal treatment influence treatment outcome, many more field E. coli and Salmonella spp isolates and clinical re­ studies must be completed to validate the currently as­ sponse to antimicrobial treatment,8•18 although these signed antibiotic break points for pathogens causing clini­ studies did not differentiate enterotoxigenic and non­ cal mastitis. enterotoxigenic strains of E. coli, and also failed to sta-

SEPTEMBER, 2004 14 tistically test the association. It is clear that we ur­ Antemortem culture of the bovine respiratory tract gently need studies documenting the antimicrobial has used guarded nasopharyngeal swabs, guarded tra­ susceptibility of E. coli isolates from the small intes­ cheal swabs, bronchoalveolar lavage (BAL) and tran­ tine of untreated calves, based on achievable drug con­ stracheal washes. Currently, endoscopic assisted BAL centrations and dosage regimens. Until these data are and transtracheal wash provide gold standard methods available, it appears that antimicrobial efficacy is best for obtaining a lower respiratory tract culture in live evaluated by the clinical response of a number of calves cattle. Unfortunately, both techniques are rarely per­ to treatment, with calves randomly assigned to treat­ formed because they are time consuming and require ment groups. The current evidence does not support specific training and appropriate restraint of the ani­ performing in vitro antimicrobial susceptibility test­ mal or expensive and fragile equipment. Nasopharyn­ ing on fecal E. coli isolates. geal swabs are commonly used to collect samples from Agar diffusion has more clinical relevance for pre­ cattle in the field because the technique is rapid and dicting the clinical response to antimicrobial treatment inexpensive;37 however, nasal swabs should not be used when applied to blood isolates than fecal isolates. This to identify the presence oflower respiratory pathogens is because the MIC breakpoints are based on achiev­ in individual cattle.1 able antimicrobial concentrations in human plasma and A nasopharyngeal swab is obtained by clearing the

MIC90 values for human E. coli isolates, which provide muzzle of accumulated secretions using a disposable a reasonable approximation to achievable MIC values towel, and then advancing a guarded swab (20 cm in in calf plasma and MIC90 values for bovine E. coli iso­ length) through the nose to sample the caudal nasal lates. Unfortunately, susceptibility results are not avail­ passage.16 A tracheal swab is obtained by restraining able for at least 48 hours, and very few studies have cattle using a head gate and two halters to minimize documented the antimicrobial susceptibility of blood lateral movement, and by applying nose tongs to elevate isolates in calves with diarrhea, and one of these stud­ and extend the head. Alaryngoscope or speculum is in­ ies observed a clinically significant year to year differ­ serted into the buccal cavity, the dorsum of the tongue ence in the results of susceptibility testing that probably is depressed, the larynx is visualized, and a guarded reflected different antimicrobial administration proto­ equine uterine swab (83 cm length) is advanced through 16 0 cols on the farm.17 the larynx and a tracheal swab obtained. Neither the "d (t) nasal pharyngeal or tracheal swab cultures the lower i::::l Pneumonia respiratory tract of adult cattle, although a long tra­ ~ (") The most important determinant of antimicrobial cheal swab (83 cm) can theoretically reach the lower (t) rn efficacy in treating pneumonia is obtaining an effective respiratory tract in recently weaned cattle. Studies have rn &. antimicrobial concentration at the site of infection, which clearly shown that bacterial populations in the upper rn is the lower respiratory tract. This is a different require­ respiratory tract differ from those in the lower respira­ q & ment to that for metaphylaxis, where the goal is to mini­ tory tract,1 •37 which is the site of infection in cattle with s. mize or prevent proliferation ofMannheimia hemolytica pneumonia. It is currently unknown whether the long 5· p in the upper and lower respiratory tract. tracheal swab provides a similar sample to that obtained Antimicrobial susceptibility testing has frequently from lung parenchyma, transtracheal wash, or BAL. been recommended to guide the treatment of respira­ However, 30% (8/27) of Mannheimia hemolytica isolated tory disease in cattle. The utility of periodic susceptibil­ from both nasal and tracheal swabs in the same animal ity testing to guide treatment decisions on feedlots has differ on the basis of ribotyping, and 37% (10/27) differ not been verified and is questionable, given that strains on the results of antimicrobial susceptibility testing.16 of Mannheimia hemolytica in a single outbreak of bo­ These findings clearly indicate that nasal and tracheal vine respiratory disease vary between and within an swabs culture different bacterial populations animal.25 A major difficulty with susceptibility testing The blind BAL technique requires similar restraint is obtaining a representative culture of bacteria from to that for the tracheal swab, but a longer flexible ster­ the lower respiratory tract of cattle with pneumonia. ile tube (100 to 240 cm in length) is passed through the The gold standard method is culturing affected outer sleeve of a guarded equine uterine swab (after the anteroventral lung parenchyma at necropsy;37 however, swab has been removed) positioned in the tracheal lu­ cattle dying of pneumonia have usually been treated men. The long flexible tube is then advanced deep into with antimicrobial agents, which increases the percent­ the lower respiratory tract so that the tube wedges into age ofresistant isolates.2•20•22 Necropsy sampling is there­ a bronchus.37 Approximately 50 mL of sterile buffer is fore strongly biased towards treatment failures. then injected through the tube and aspirated, provid­ Practical methods for obtaining a representative cul­ ing a lavage of the bronchoalveolar unit. In the endo­ ture of the lower respiratory tract bacteria in untreated scopic assisted BAL technique, an endoscope is passed cattle are therefore needed. through the nasal passages and down the trachea to

15 THE AABP PROCEEDINGS-VOL. 37 the carina, and the right apical lobe bronchus identified priately validated, and because bacteria are usually for sampling.1 The transtracheal wash technique starts cultured from nasal or tracheal swabs, which have not with aseptic preparation of the skin in the ventral mid been shown to represent lower respiratory tract bacte­ cervical region, percutaneous passage of a sterile can­ rial populations, or cultured from dead animals, which nula into the tracheal lumen, and passage of a sterile represent treatment failures. flexible catheter through the cannula towards the ca­ rina. Alternatively, the sterile flexible tube is advanced Conclusions through the outer sleeve of a guarded equine uterine swab in a similar manner to BAL. 25 Once positioned in Antimicrobial susceptibility testing costs up to $20 the region of the carina, sterile phosphate buffered sa­ per test. Because the cattle industry is economically line in injected through the catheter, reaspirated and driven, any diagnostic test should be validated, have the recovered fluid submitted for bacterial culture. The appropriate sensitivity and specificity, and an accept­ transtracheal wash and endoscopic assisted BAL cul­ able economic return on the cost of testing before it can ture the anteroventral section of the lung, which is the be routinely recommended. Antimicrobial susceptibil­ predilection site for bacterial pneumonia. Transtracheal ity testing has not been adequately validated for the wash is theoretically preferable to a blind BAL proce­ treatment of masti tis, calf diarrhea and pneumonia. The dure because the latter technique cultures the diaphrag­ use of susceptibility testing to guide treatment decisions matic lung lobes, which are affected only in severe cases for individual cattle is therefore not recommended. How­ of pneumonia. ever, because the results of susceptibility testing are Studies have conclusively shown that nasal swabs repeatable, the results of population susceptibility test­ are inaccurate indicators of lung parenchymal Myco­ ing do provide useful information on the development plasma species in cattle.37 It is therefore difficult to in­ or loss of antibiotic resistance characteristics for patho­ terpret the results of most pneumonia studies where gens in a population over time. In calf diarrhea, bacte­ nasal SW8;bS are used to monitor susceptibility. For in­ rial isolates for susceptibility testing should be obtained stance, the findings in one study that a poor correlation from the small intestine of untreated calves and con­ existed between the MIC values for Mannhemia firmed to be enterotoxigenic strains of E. coli. In cattle hemolytica and response to treatment with penicillin, with pneumonia, bacterial isolates for susceptibility test­ oxytetracycline, or trimethoprim-sulfadoxine suscepti­ ing should be obtained from untreated animals using a bility23 may not have represented the findings if culture transtracheal wash or endoscopic assisted BAL, or pos­ samples had been obtained using a transtracheal wash sibly a tracheal swab in low body weight cattle. Finally, or endoscopic assisted BAL In other words, therapeutic because zone diameter is the measured variable in the strategies should not be based on the antimicrobial sus­ agar diffusion test, it is recommended that studies us­ ceptibilities of nasal or tracheal bacterial isolates, be­ ing the results of agar diffusion (Kirby-Bauer test) to cause these have not been validated as representing describe antimicrobial susceptibility report the mea­ lower respiratory tract infection. sured value (zone diameter) as well as the interpreta­ We do not currently have adequate databases ofin tion of the zone diameter (susceptible, intermediate, vitro MIC values for bacterial isolates from cattle with resistant). This recommendation is made because the pneumonia, because almost all isolates were obtained interpretive criteria for agar diffusion and MIC from nasal swabs oflive animals29 or lung parenchyma breakpoints will obviously change when the breakpoints of dead animals that represented treatment failures. 29•42 are appropriately validated. In some studies the source of the isolates was not stated.32 All the pivotal studies for FDA approval of References tilmicosin, spectinomycin and enrofloxacin used pre­ treatment nasal swabs and necropsy lung swabs (treat­ 1. Allen JW, Viel L, Bateman KG, et al: The microbial flora of the ment failures) to characterize the susceptibility profile, respiratory tract in feedlot calves: associations between nasopharyn­ geal and bronchoalveolar lavage cultures. Can J Vet Res 55:341-346, whereas the pivotal studies for ceftiofur and florfenicol 1991. utilized pretreatment nasal swabs, transtracheal washes 2. Allen JW, Viel L, Bateman KG, et al: Changes in the bacterial flora and lung swabs (treatment failures) obtained at necropsy of the upper and lower respiratory tracts and bronchoalveolar lavage (www.fda.gov/cvm/efoi). The use of nasal swabs for sus­ differential cell counts in feedlot calves treated for respiratory dis­ eases. Can J Vet Res 56:177-183, 1992. ceptibility testing therefore casts doubt on the accuracy 3. Ames TR, Larson VL, Stowe CM: Oxytetracycline concentrations of NCC LS recommended breakpoints for antibiotics used in healthy and diseased calves. Am J Vet Res 44:1354-1357, 1983. to treat bacterial pneumonia in cattle. In summary, the 4. Anonymous. NCCLS document M31-A2. Performance standards results of susceptibility testing should not be currently for antimicrobial disk and dilution susceptibility tests for bacteria isolated from animals; approved standard, ed 2. Wayne NJ, National used to guide treatment decisions in cattle with pneu­ Committee for Clinical Laboratory Standards, 2002. monia because the break points have not been appro-

SEPTEMBER, 2004 16 5. Anonymous. Terminology relating to methods for the determina­ 25. Murphy GL, Robinson LD, Burrows GE: Restriction endonuclease tion of susceptibility of bacteria to antimicrobial agents. Clinical Mi­ analysis and ribotyping differentiate Pasteurella haemolytica sero­ crobiology and Infection 6:503-508, 2000. type Al isolates from cattle within a feedlot. J Clin Microbiol 31:2303- 6. Anonymous. Determination of antimicrobial susceptibility test 2308, 1993. (Q) breakpoints. Clinical Microbiology and Infection 6:570-572, 2000. 26. Myllys V, Louhi M, Ali-Vehmas T: Comparison of Pencillin-G sus­ ceptibility testing methods of Staphylococci isolated from bovine mas­ n 7. Bauer AW, Kirby KM, Sheris JC, et al: Antibiotic susceptibility test­ 0 ing by a standard disc method. Am J Clin Pathol 45:493-496, 1966. titis. J Vet Med B 39:723-731, 1992. "O '< 8. Boyd JW, Baker JR, Leyland A, et al: Neonatal diarrhea in calves. 27. Mylrea PJ: Passage of antibiotics through the digestive tract of ....'""I . Vet Rec 95:310-313, 1974. normal and scouring calves and their effect upon the bacterial flora. {IQ 9. Burrows GE: Effects of experimentally induced Pasteurella Res Vet Sci 9:5-13, 1968. g' haemolytica pneumonia on the pharmacokinetics of erythromycin in 28. Owens WE, Ray CH, Watts JL, et al: Comparison of success of ► the calf. Am J Vet Res 46:798-803, 1985. antibiotic therapy during lactation and results of antimicrobial sus­ ~ 10. Bywater RJ, Palmer GH, Wanstall SA: Discrepancy between anti­ ceptibility tests for bovine mastitis. J Dairy Sci 80:313-317, 1997. ....'""I . biotic (amoxycillin) resistance in vitro and efficacy in calf diarrhea. 29. Post KW, Cole NA, Raleigh RH: In vitro antimicrobial susceptibil­ (') Vet Rec 102:150-151, 1978. ity of Pasteurella haemolytica and Pasteurella multocida recovered § 11. Chin AC, Morck DW, Merrill JK, et al: Anti-inflammatory benefits from cattle with bovine respiratory disease complex. J Vet Diagn In­ Cfl► Cfl of tilmicosin in calves with Pasteurella haemolytica-infected lungs. vest 3:124-126, 1991. 0 Am J Vet Res 59:765-771, 1998. 30. Pyorala S, Syvajarvi J: Bovine acute mastitis. Part II. Effect of (') 12. Clark JG, Adams CJ, Addis DG, et al: Oxytetracycline blood se­ mastitis pathogen, initial inflammatory reaction and therapy on the ~.....- rum levels in healthy, pneumonic, and recovered cattle. Proc Am Assoc outcome of the disease. J Vet Med B 34:629-639, 1987. 0 Bou Pract 9:142-147, 1976. 31. Salmon SA, Watts JL, Yancey RJ: In vitro activity of ceftiofur and :::1 0 13. Constable PD: Antimicrobial use in the treatment of calf diar­ its primary metabolite, desfuroylceftiofur, against organisms of vet­ H; rhea. J Vet Int Med 18:8-17, 2004. erinary importance. J Vet Diagn Invest 8:332-336, 1996. to 0 14. Constable PD, Morin DE: Use of antimicrobial susceptibility test­ 32. Shryock TR, White DW, Staples JM, et al: Minimum inhibitory < ing of bacterial pathogens isolated from the milk of dairy cows with concentration breakpoints and disk diffusion inhibitory zone inter­ s· clinical mastitis to predict response to treatment with cephapirin and pretive criteria for tilmicosin susceptibility testing against Pasteurella (1) oxytetracycline. J Am Vet Med Assoc 221:103-108, 2002. spp. associated bovine respiratory disease. J Vet Diagn Invest 8:337- ~ 15. Constable PD, Morin DE: Treatment of clinical mastitis. Using 344, 1996...... (') antimicrobial susceptibility profiles for treatment decisions. Vet Clin 33. Smith HW: Further observations on the effect of chemotherapy on ...... North Am Food Animal Pract, 19:139-155, 2003 the presence of drug-resistant Bacterium coli in the intestinal tract of §i' 16. Derosa DC, Mechor GD, Staats JJ, et al: Comparison of Pasteurella calves. Vet Rec 70:575-580, 1958. (1) '""I spp. simultaneously isolated from nasal and transtracheal swabs from 34. Smith HW: Observations on the etiology of neonatal diarrhoea Cfl cattle with clinical signs of bovine respiratory disease. J Clin Micro (scours) in calves. J Path Bact 84:147-168, 1962. 0 "O 38:327-332, 2000. 35. Smith HW, Crabb WE. The typing of E. coli by bacteriophage, its (1) 17. Fecteau G, Van Metre DC, Pare J, et al: Bacteriological culture of application to the study oftheE. coli population of the intestinal tract :::1 blood from critically ill neonatal calves. Can Vet J 38:95-100, 1997. of healthy calves and of calves suffering from white scours. J gen ~ (') 18. Glantz PJ, Kradel DC, Seward SA: Escherichia coli and Salmo­ Microbiol 15:556-574, 1956. (1) Cfl nella newport in calves: efficacy of prophylactic and therapeutic treat­ 36. Smith HW, Crabb WE: The sensitivity to chemotherapeutic agents Cfl ment. Vet Med Small Anim Clin 69:77-82, 1974. of a further series of strains of Bacterium coli from cases of white &. 19. Groothuis DG, van MiertASJPAM, Nouws JFM: Effects of experi­ scours: the relationship between sensitivity tests and response to treat­ .....Cfl mental Escherichia coli endotoxemia on ampicillin:amoxycillin blood ment. Vet Rec 68:274-277, 1956. ;.:'""I levels after oral and parenteral administration in calves. J Vet Pharm 37. Thomas A, Dizier I, TrolinA, et al: Comparison of sampling proce­ Therapeutics 1:81-84, 1978. dures for isolating pulmonary mycoplasmas in cattle. Vet Res a..... 0 20. Hjerpe CA, Routen TA: Practical and theoretical considerations Communic 26:333-339, 2002. p concerning treatment of bacterial pneumonia in feedlot cattle, with 38. Thornsberry C, Burton PJ, Yee YC, et al: The activity of a combi­ special reference to antimicrobic therapy. Proc Am Assoc Bou Pract nation of penicillin and novobiocin against bovine mastitis pathogens: 9:142-147, 1976. development of a disk diffusion test. J Dairy Sci 80:413-421, 1997. 21. Lakritz J , Tyler JW, MarshAE, et al: Tilmicosin reduces lipopolysac­ 39. Thornsberry C, Marler JK, Watts JL, et al: Activity ofpirlimycin charide-stimulated bovine alveolar macrophase prostaglandin E2 pro­ against pathogens from cows with mastitis and recommendations for duction via a mechanism involving phospholipases. Vet Therapeut disk diffusion tests.AntimicrobAgents Chemother 97:1122-1126, 1993. 3:7-21, 2002. 40. Vuye A, Pijck J, Soep H: In vitro activity ofBL-S640 against gram­ 22. Martin SW, Meek AH, Curtis RA: Antimicrobial use in feedlot negative bacilli and Staphylococcus aureus compared with activity of calves: its association with culture rates and antimicrobial suscepti­ four other semisynthetic cephalosporins. Antimicrob Agents Chemother bility. Can J Comp Med 47:6-10, 1983. 9:422-432, 1976. 23. Mechor GD, Jim GK, Janzen ED: Comparison of penicillin, ox­ 41. Watts JL, Salmon SA, Yancey RJ, et al: Antimicrobial susceptibil­ ytetracycline, and trimethoprim-sulfadoxine in the treatment of acute ity of microorganisms isolated from the mammary glands of dairy undifferentiated bovine respiratory disease. Can Vet J 29:438-443, heifers. J Dairy Sci 78:1637-1648, 1995. 1988. 42. Watts JL, Yancey RJ, Salmon SA, et al: A 4 year survey of antimi­ 24. Morin DE, Shanks RD, McCoy GC : Comparison of antibiotic ad­ crobial susceptibility trends for isolates from cattle with Bovine res­ ministration in conjunction with supportive measures versus support­ piratory disease in North America. J Clin Micro 32:725-731, 1994. ive measures alone for treatment of dairy cows with clinical mastitis. J Am Vet Med Assoc 213:676-684, 1998.

17 TH E AABP PROCEEDINGS-VOL. 37