WHICH ANTIMICROBIAL SHOULD I USE? OPTIMIZING ANTIBACTERIAL THERAPY FOR SMALL ANIMALS USING THE PROFESSIONAL FLEXIBLE LABEL Dawn M. Boothe, DVM, PhD, DACVIM, DACVCP Department of Veterinary Physiology and Pharmacology, College of Veterinary Medicine, Texas A&M University, College Station, Texas, 77843-4466 dboothe@cvm.tamu.edu www.cvm.tamu.edu/vcpl 409 845 9368 INTRODUCTION The advent of professional flexible labeling (PFL) has led to an increased awareness of the need for individualizing antimicrobial therapy (Martinez 1995; Farho 1995; Sundlof 1995; Martinez 1996). . Until recently, dosing regimens (dose and interval) were "fixed" in that only one dose and one interval were approved (labeled) for the animal. Fixed labels tend to be restrictive to the practice of medicine because each is applied to the patient, regardless of its clinical status or the severity or location of infection. Generally lacking on fixed labels is information that might guide modification of dosing regimen to facilitate therapeutic efficacy in the presence of host and microbial factors that complicate infection. In contrast, the PFL not only provides several dosing regimens such that more appropriate regimens can be selected for individual patients, but also is accompanied by information that can facilitate selection of the regimen that is most appropriate. However, to take advantage of the new PFL, one should understand the complex relationship between pharmacokinetics (Martinez 1998a-c) and pharmacodynamics, that is the identity of the infecting microbe, the pathophysiology of infection as it has affected the patient and the clinical pharmacokinetics of the chosen drug. General Considerations. Antimicrobial therapy provides some unique considerations for individualization of dosing regimens. Some of these considerations complicate antimicrobial therapy whereas others facilitate therapy. Confounding considerations include the following. First, unlike other drugs, the target of antimicrobial therapy is not a host tissue, but a microbe simply located in host tissues. Unfortunately, the identify of the microbe often is not known, but rather is based on historical data (ie, most likely infecting microbe) and drug selection is thus empirical (Shikawa 1984). Second, unlike most drugs (eg, cardiac, respiratory or gastrointestinal), antimicrobials potentially must reach any tissue in the body. Infections can occur anywhere, including the most difficult to reach tissues such as the brain (Lefrock 1984) . Third, microbial and host factors at the site of infection often effectively decrease the amount of active drug at the site, leading to therapeutic failure. Confounding factors at the site include accumulation of protein (Craig 1989) and inflammatory debris (Brumbaugh 1991). Inflammatory debris can be problematic because it presents a barrier to passive diffusion, the major drug movement, and binds to and inactivates some drugs. Additional, oxygen tension can be reduced at the site of infection (rendering aminoglycosides ineffective and impairing host white blood cell phagocytic activity), and pH can be increased (ionizing and rendering less penetrable weak bases such as aminoglycosides and fluorinated quinolones) (Neu 1994). With time, deposition of fibrous tissue will further preclude drug movement to the site of infection. Fourth, unlike other drug therapy (excepting anticancer drugs), antibiotics are intended to kill the target (microbe). Thus, drug concentrations must be sufficiently high to penetrate and kill the organism without harming the host. In addition, host immune response often must be competent for selected antimicrobials to be effective. A fifth confounder is the ability of microbes to protect themselves through mechanisms which not only might damage the host, but also may result in antimicrobial resistance. Among the mechanisms of microbe resistance is the ability of some organisms to survive phagocytosis. In such instances, antimicrobials must be able to penetrate white blood cells in order to be effective. Failure to achieve sufficient drug concentrations of the appropriate drug at the site of infection will facilitate antimicrobial resistance not only to the chosen drug, but probably to other antimicrobials as well. Several considerations facilitate antimicrobial therapy (Neu 1994). First, microbes are prokaryotes whereas the patient is eucaryote. Thus, the targets of antimicrobial drugs generally are sufficiently different from host tissues that the drugs tend to be safe. Indeed, the mechanisms of toxicity of antimicrobials are often unrelated to their mechanism of antimicrobial action. Second, and most importantly, through culture and susceptibility data, not only can the specific microorganism(s) be identified, but information regarding the susceptibility of the organism to the drug of interest (that is, drug efficacy) can provide guidelines for dose modification. Professional flexible labels contain both pharmacokinetic and pharmacodynamic data useful to the design of a dosing regimen for the individual patient (Martinez 1996). Pharmacodynamic data includes information regarding the minimum inhibitory concentration (MIC) of the drug for organisms in the spectrum of the antimicrobial. Pharmacokinetic data includes information regarding the disposition of the drug in the target species. The design of a dosing regimen should begin with identification of a drug to which the organism is susceptible. Selection ideally is based on culture and susceptibility data; if not available, the the drug is empirically (based on historical evidence of most likely infecting microbe) (Yoshikawa 1984). However, subsequent modification of the regimen should be based on consideration of drug, microbial and host factors, including changes at the site of infection that might preclude antimicrobial efficacy. Resistance to Antimicrobial Agents. The role of resistance in therapeutic failure of antimicrobials is not uncommon. The ability of organism to develop resistance varies with the strain. The mechanisms of bacterial resistance vary and involve changes in cell wall structures, proteins (ie, penicillin-binding proteins), or enzymes, development of enzymes which destroy antimicrobials (ie, ß-lactamases which destroy penicillins), and changes in intracellular transport proteins (tetracyclines), metabolic pathways (sulfonamides) or binding sites (ie, on ribosomes as for aminoglycosides) for antibiotics. Plasmid-mediated resistance in gram negative organisms is common, can develop rapidly and can be transmitted between species. A single transfer of plasmid genetic material from a bacterial donor can result in antimicrobial resistance to up to 7 antimicrobials in the recipient. The occurrence of resistance is facilitated by : generation of subtherapeutic drug concentrations such as might occur when using an inappropriate dosing regimen, or non-IV administration; failure to base antibiotic selection on culture and susceptibility results; indiscriminant use of broad spectrum antibiotics. Bacterial resistance has been decreased by: "protecting" the antibiotic (eg, clavulanic acid); modifying the compound so that it is more difficult to destroy (eg, amikacin compared to gentocin); use of lipid soluble compounds that are more able to achieve effective concentrations (eg, doxycycline compared to other tetracyclines); use of beta-lactams in combination with other drugs to enhance cellular penetration; and use of combination antimicrobials that target the same organism. PHARMACODYNAMIC DATA Susceptibility Data Provided on the PFL. Microbiological data on a PFL will reflect the results of agar gel disc diffusion (eg, Kirby Bauer; Acar 1996) or tube dilution methods (Amsterdam 1996) of susceptibility testing. Disc diffusion data will include zone diameters indicative of organism susceptibility (S) or resistance (R). For example, current zone diameters for enrofloxacin at a dose of 2.5 mg/kg are > 20 mm = S (susceptible), < 16 = R (resistant), and between 17-19 mm = I (intermediate). Tube dilution susceptibility will be presented as minimum inhibitory concentration (MIC) data (Figure 1). The MIC of an organism for a drug refers to the minimum amount of drug necessary to inhibit visible growth of an organism using standardized culturing methods as guided by the National Committee for Clinical Laboratory Standards (NCCLS) of the Center for Disease Control (CDC). The MIC data on a PFL may be presented in several forms, including the range of MIC for susceptible organisms, the MIC50 and the MIC90. The latter data are the MIC at which 50% and 90%, respectively of the isolates (by genus and species) are inhibited (not killed ). However, the MIC50 and MIC90 ideally are based on a large number of microorganisms (ideally, more than 100). For disc diffusion data, zone diameters correlate with MIC data. For either method of susceptibility testing, the likelihood of a drug being effective can be based on whether or not the recommended dose on the label is likely to generate plasma drug concentrations (PDC) that equal or surpass the MIC of the infecting organism. The breakpoint MIC (MICBP) of a drug might be considered a "yardstick" against which MIC for an infecting organism can be compared. Simplistically, the MICBP is the MIC below which will fall the MIC of susceptible organisms. The MICBP should approximate the peak PDC that will be achieved using the recommended dose of the drug and should be correlated with clinical response to the selected antimicrobial (Amsterdam 1996). If the MIC of the organism is sufficiently lower than the MICBP , the organism is considered susceptible (S). If the MIC of the organism equals or surpasses the MICBP, the organism is considered resistant (R). Breakpoint MICs most appropriately are approved by the NCCLS and should be the same for all laboratories because it is dependent on the microbe and the targeted animal species. Hence, MICBP will not be available for all approved veterinary drugs, even those with PFL. Care should be taken in assuming that all drugs in the same class (eg, all veterinary fluorinated quinolones) are equally susceptible to a drug. For example, susceptibility to enrofloxacin should not be interpreted as susceptibility to other veterinary fluorinated quinolones, particularly with organisms that approach the MICBP. One precaution must be taken into account when using MICBP on PFL to guide therapy. The MICBP has been based on a single dose, yet PFL offer a range of doses. For example, enrofloxacin originally was approved in the US with a fixed dosing regimen and a MICBP based on a dose of 2.5 mg/kg. Yet, dosing regimens range from 5 to 20 mg/kg. A PFL since has been approved with doses that range up to 20 mg/kg once daily. The original MICBP does not reflect the higher dose and organisms identified as "R" might actually be susceptible at the higher dose. However, this summer, a new breakpoint MIC reflecting the PFL was identified for enrofloxacin by the NCCLS: "S" = MIC < 1 µg/ml, "R" = >4 µg/ml, and "F" = MIC between 1 and 4 µg/ml, indicating the need to use higher doses as suggested on the PFL. New zone diameters similarly have been identified: > 23 mm = S, < 16 = R and 18-22 = F. Veterinary microbiology laboratories should be encouraged by practitioners to implement changes in interpretive susceptibility data which reflect the flexibility in dosing regimens. Figure 1. Tube dilution data will be presented as an MIC. The MIC of the organism is compared to the breakpoint MIC for the drug (in parenthesis) to determine the S,I or R designation. Initial Dose Selection. The more information that is known regarding the organism causing infection in the patient, the more useful microbiologic data on the PFL. The most benefit is gained if the MIC of the infecting organism is known. Veterinary microbiology laboratories are increasingly adapting this method of susceptibility testing. The MIC of the infecting microorganism can be compared to the MICBP in order to appreciate the relative susceptibility of the organism to the drug; this in turn may guide selection of the most appropriate dose. The lower the MIC compared to the MICBP, the lower the dose that might be used. For example, for enrofloxacin, an organism with an MIC of 0.06 µg/ml is obviously more likely to succumb to a dose of 5 mg/kg compared to an organism with an MIC of 1 µg/ml. Treatment of the latter should be based on an initial dose selection in the mid or high end of the recommended dosing range (eg, 7.5 to 20 mg/kg), depending on other factors (see further discussion). Although less ideal, the same approach can be used when basing dose selection on agar gel diffusion data or when the organism is suspected but not known. For disc diffusion data, only "S", "I" and "R" designations are given; no evidence of relative susceptibility is provided. However, the MIC90 data on the PFL can help guide initial selection of a dosing regimen when only disc diffusion data is available. The low end of a dosing range may be appropriate for Staphylococcus intermediaus whose most recently (NCCLS approved) MIC90 is. In contrast, the higher end of the dosing regimen is indicated if the organism is Pseudomonas aeruginosa whose most recently established MIC90 is 1 µg/ml (Table 1). The same is true for both organisms when treating with marbofloxacin (Table 1). For orbifloxacin, the lower dose might be appropriate for; however, even the higher dose is not likely to be effective against Pseudomonas aeruginosa (Table 1). Relationship between MIC and tissue Drug Concentrations The bridging of pharmacodynamic (susceptibility) data with pharmacokinetic data should begin with an appreciation of the relationship between the MIC established by in vitro testing and the concentrations of drug achieved in the patient at the site of infection when the drug is administered at the labeled dose. Two drug related issues should be considered as a dose is selected. Bactericidal versus bacteriostatic drugs. "Bactericidal" drugs are defined as drugs whose MIC is very close to the minimum bactericidal concentration (MBC) (Alexander 1996). Bactericidal effects can be related to mechanism of action ; drugs which inhibit ribosomes and thus protein synthesis are less likely to be bactericidal. Both MIC and MBC reflect in vitro data; however, concentrations which effectively kill the infecting microbe are more likely to be achieved safely in the patient with a bactericidal drug compared to a bacteriostatic drug. Thus, use of selected drugs, such as the fluorinated quinolones, beta-lactams, aminoglycosides and potentiated sulfonamides, often occurs because the practitioner is confident of efficacy. However, the bactericidal effects of the drug depend on achieving sufficient concentration of the drug at the antimicrobial target (Levison 1989). The effects of such drugs can easily be rendered "bacteriostatic" should insufficient concentrations be reached. Post-antibiotic effect. The elimination half-life of many antimicrobials ranges from 1 to 4 hours. Most dosing intervals range from 8 to 12 hours. For many drugs, concentrations in plasma or tissue are nondetectable at the end of the dosing interval. Yet, antimicrobial efficacy may not be impaired for some drugs. Persistence of antimicrobial effects after brief exposure to (or the lack of detectable concentrations of) an antimicrobial has been termed the post-antibiotic effect (PAE) (Brown 1987, Crain 1996, Levison 1989, Spivey 1992, John 1998, Wetzstein 1994). The PAE can prolong the dosing interval (Brown 1987) and is therapeutically important for some antimicrobials against some organisms. For some drugs (eg, fluorinated quinolones and aminoglycosides), the duration of the PAE (and thus antibiotic efficacy) is concentration dependent and is maximized by a large PDC:MIC or inhibitory quotient (the ratio of PDC:MIC) (Moore 1987, Schentag 1991b, Volgeman 1988). In contrast, efficacy of beta-lactams and most bacteriostatic drugs is considered time (time of exposure) dependent. For such drugs, PDC should remain above the MIC during the majority of the dosing. This variability can impact the dosing regimen for a drug as is exemplified by comparing beta-lactams (interval or time dependent) with aminoglycoside and fluorinated quinolones (dose or concentration dependent) antimicrobials. Efficacy of the concentration dependent drugs is enhanced by administering a high dose which maximizes the inhibitory quotient. In contrast, efficacy of beta-lactam or bacteriostatic antimicrobials is enhanced by using shorter dosing intervals, although a dose increase may be necessary in some instances in order to surpass the MIC. Thus, using a dose that is too low is particularly detrimental with fluorinated quinolones, whereas prolonging the dosing interval should be avoided for beta-lactams. It is important to note that reported relationships between PDC, MIC and therapeutic efficacy (and the PAE) is based on in vitro data; additionally, the relationships vary with drugs and organisms. More recent literature has proposed the area under the inhibitory curve (AUIC) as a better predictor of antimicrobial efficacy (Schentag 1991a,b). This variable is derived by dividing the area under the PDC versus time curve by the MIC of the infecting organism; as such, both peak concentrations and elimination half-life are relevant to drug efficacy. Pharmacokinetic Data The inclusion of data describing the disposition of drug in the target species can be a powerful tool for guiding design of a dosing regimen. Interpreting this data represents the first step in applying pharmacodynamic data to the individual patient. However, the data can also be misleading if not intepreted in the proper perspective. Maximum PDC. Most infections occur in extracellular (interstitial) fluid (ECF). For most tissues, capillaries are not fenestrated and little or no barrier is presented to passive drug movement from plasma to the site of infection. For these "simple" infections, PDC can be compared to MIC (or MIC90) of the infecting organism in order to evaluate a dose (Levin 1984). For PFL, generally PDC will be offered only for the low end of the dosing regimen. However, peak PDC achieved at higher doses generally can be extrapolated by proportionally increasing PDC to the dose increase (Boothe 1999). Volume of Distribution. For problematic infections, peak PDC may not be the most relevant pharmacokinetic variable upon which a dose should be selected. For tissues with non fenestrated capillaries (eg, prostate, brain, cerebral spinal fluid, prostate) or in tissues with marked inflammatory response (see below) drug distribution can be impaired by the presence of lipid barriers (Bergan 1981, Lefrock 1984). The cell membrane is an additional barrier to penetration for intracellular infections. In such instances, the penetrability of the drug in various body tissues may most appropriately reflect antimicrobial concentrations at the site of infection. Several indicators of drug penetrability may be presented on a PFL. The volume of distribution (Vd) is a pharmakokinetic variable used to design a determine doses. This theoretical volume simplistically is the tissue which will dilute drug which enters the blood. Dose is directly proportional to Vd. In fact, if the Vd and the target concentration of the drug in the patient are known, a dose can be calculated (Dose = Vd X targeted PDC). If the desired targeted concentrations of the drug is equal to the MIC of the infecting organism, then the dose in the patient = Vd * MIC of the organism. For example, for treatment of Pseudomonas aeruginosa with enrofloxacin (Vd = 2.7 l/kg), a dose can be calculated based on an MIC90 of 1 µg/ml : dose = 2.7 l/kg * 1 mg/l = 2.7 mg/kg. However, this dose will simply achieve the MIC in plasma and may not be sufficient for efficacy, particularly for concentration dependent drugs. Drugs which are water soluble and distribute only to extracellular fluid (beta-lactamas, aminoglycosides) are characterized by a Vd that tends to be less than 0.3 L/kg whereas drugs which are able to distribute to total body water have a volume which approximates 0.6 l/kg (eg, fluorinated quinolones, macrolides, selected sulfonamides, lincosamides. A Vd greater than 0.6 l/kg indicates accumulation of the drug in tissues. Although Vd is useful for evaluating the ability of a drug to move passively through lipid soluble membranes, it provides no information as to which tissues the drug moves. Thus, using Vd as a basis for selecting one drug over another must be done with caution. Certainly a drug with a Vd of 0.6 l/kg is more likely to penetrate cell membranes compared to a drug with a Vd of 0.3 L/kg. A larger VD may suggest better lipophilicity and thus penetration of selected tissues, including inflammatory debris. However, for drugs with Vd > 0.6 l/kg, differences in Vd among drugs may simply reflect differences in accumulation (retention) in one tissue compared to another. Comparison of tissue drug concentrations may be more appropriate than comparison of Vd for such drugs. Tissue drug concentrations can be compared to the MIC of the organism cultured from the patient, or the MIC90 (Table 1). However, tissue data accompanying package inserts generally is based on tissue homogenates, which includes intracellular as well as extracellular drug concentrations, yet intracellular drug may not necessarily be available to bacteria located in ECF. Exceptions include tissues of elimination (eg, bile and urine). For such tissues, antimicrobial drugs will achieve much higher concentrations compared to plasma and if infection is in the organ of elimination, MIC data based on PDC may markedly underestimate antimicrobial efficacy. As a class, fluorinated quinolones are accumulated in phagocytic white blood cells (Easmon 1985, Tulkens 1990); of the veterinary fluorinated quinolones, this effect has been documented for enrofloxacin (Hawkins 1996, Boeckh 1999) and marbofloxacin (unpublished data). Accumulation in white blood cells will not only increase efficacy against intracelluar bacteria, but probably increases the concentration of drug at the site of inflammation (infection) (Boothe 1999b). Selection of the dose should take into account potential accumulation of drug in the target tissue. Miscellaneous information. Other pharmacokinetic data provided on the PFL will be variably useful, depending on the circumstances surrounding infection. Elimination half-life will be particularly useful for time-dependent antimicrobials and should provide the basis for interval selection for such drugs. Plasma drug concentrations will decline by 50% if the dosing interval equals the elimination half-life. Selection of a dosing interval should be based on the distance between PDC from the MIC of the organism and the amount of time that can elapse before PDC decrease below the MIC. Although a physiologically relevant variable, clearance does not provide much pharmacokinetic guidance for design of a dosing regimen. However, the route of clearance, can provide critical information in the animal whose organs of clearance for the antimicrobial is dysfunctional (Lesars 1984, Riviere 1984 ). The PFL will not be all inclusive; information relevant to dose modification may be absent for some drugs, in part because information on the label must be approved by the Food and Drug Administration. Thus, clinicians should continue to supplement label information with current literature when appropriate. For example, formation of active metabolite of a drug may not appear on a label. A significant amount of ciprofloxacin is formed from enrofloxacin. Although this might may impact therapeutic decision making, the information can be collected only from veterinary literature (Kung 1993, Hawkins 1996,Boothe 1999). Pathological Factors Impacting the Dosing Regimen Host Factors. Thus far, factors influencing the design of a dosing regimen has been based on information from a PFL in healthy tissues. However, host response at the site of infection also must be considered because of the potential profound effect on it may have on antimicrobial efficacy. The microenvironment of the site of infection can impair activity of many antimicrobials through several mechanisms; the impact increases with chronicity of infection. Accumulation of cellular debris associated with the inflammatory process can present a barrier to passive antibiotic distribution. Local pH becomes more acidic as degradative products such as lysosomes, nucleic acids, and other intracellular constituents released from dying white blood cells accumulate. The efficacy of weakly basic antibiotics (eg, aminoglycosides and fluorinated quinolones) can subsequently be impaired as the drugs become more ionized. Selected drugs are bound to and thus inactivated by hemoglobin and proteinaceous or other debris that accumulates with inflammation. With chronicity, deposition of fibrous tissue at the infected site will further reduce drug movement into target tissues. Low tissue oxygen tension which can accompany inflammation reduces WBC phagocytic and killing activity, which may contribute to therapeutic failure, particularly if drug concentrations at the site are not bactericidal. Low oxygen tension also may slow the growth of organisms, rendering them less susceptible to drugs whose mechanism of action is dependent on rapid microbial growth. Microbial factors. A number of microbial factors can impact therapeutic success with an antimicrobial and might lead the clinician to select a higher dose. The size of the bacterial innoculum can also influence antibiotic efficacy. The larger the bacterial innoculum at the target site, the greater the concentration (number of molecules) of antibiotic necessary to kill the organisms, and the greater the risk of antibiotic destruction by enzymes or other materials produced by microorganisms. In addition, microbes can produce materials that decrease penetration (eg,"slime" of Staphylococcus, glycocalyx of Pseudomonas, calcium containing granules of Nocardia). Host response to infection and its impact on antimicrobial therapy may vary with the organ system infected. For example, in respiratory tract infections, mucus produced by the host also can directly interfere with antimicrobial therapy. Antibiotics may bind to glycoproteins and mucus may present a barrier to passive diffusion. Antimicrobial resistance, particularly plasmid-mediated, is facilitated by insufficient drug concentrations at the site of infection (Neu 1994). Decreased porin size and increased destruction of drug are two examples of resistance commonly mediated by plasmids. This type of resistance occurs rapidly (during therapy), is transmissible among different genuses, and can occur simultaneously for multiple drugs. Previous antimicrobial therapy might be considered as a reason to pick a higher dosing regimen. Designing a Dosing Regimen for Antimicrobials Dose. A number of factors have been discussed which may impact the selection of a dose. The low end of a dosing range is more likely to be effective for first-time infections caused by non-intracellular organisms whose MIC is distant from the MICBP, MIC90 or predicted tissue concentrations, and if host response to the infection is mild. The more complex the response to infection, the more difficult the tissue is to penetrate or the closer the MIC of the organism to the MICBP, MIC90 or tissue concentration (including PDC) of the drug, the closer the selected dose should be to the higher end of the dosing range. For concentration dependent drugs, the dose ideally will be sufficient to generate at the site of infection an inhibitory quotient that is between 8 - 10. The inhibitory quotient may need to be higher in serum if the inflammatory response by the host is marked. For time dependent drugs, the dose of the antibiotic may need to be increased 2 to 3 fold to compensate for decreased antimicrobial movement into the tissue or other host factors that might decrease antimicrobial efficacy. The MIC data accompanying a culture report also can be used more for calculating a specific dose. Such an approach is appropriate for identifying the lowest effective dose for very expensive or toxic drugs or if the MIC of a chosen drug is approaching breakpoint. Recalling that Dose = [Vd] x [Targeted PDC], the targeted PDC should be based on the MIC of the infecting organism, modified for drug, host and microbial factors such that targeted PDC = MIC times a "therapeutic factor". The magnitude of the factor depends on the complexity of the infection. Enrofloxacin (MICBP < 2µg/ml) or amikacin (MICBP < 32 µg/ml) provide examples. Assume an infection has been caused in the respiratory tract caused by Eschericia coli and the MIC for enrofloxacin (Vd 2.6 l/kg) has been measured at 0.5 µg/ml and for amikacin (Vd 0.25 l/kg) at of 8 µg/ml. Both drugs are concentration dependent and thus, a higher dose should be anticipated. Assume that the infection is accompanied by marked inflammation in the alveoli and peribronchial tisseus. Enrofloxacin distributes better into bronchial fluids (tissue:serum ratio > 1) compared to amikacin (tissue : serum ration < 0.3) (Boothe, 1999c). The desired inhibitory quotient for both these drugs is 8-10 times the MIC. The therapeutic factor for amikacin should be higher than that for enrofloxacin (eg, at least 10) to account for the additional complication of poor drug penetrability, especially in the presence of a marked inflammatory response. In contrast, enrofloxacin is accumulated in WBC, and marked inflammation is likely to facilitate drug concentrations at the site. Thus, an inhibitory quotient of less than 8 might be reasonable. However, 8 will be used in this calculation . The targeted PDC for enrofloxacin then would be 0.5 µg/ml X 8, or 4 µg/ml. The dose would be 2.6 l/kg X 4 µg/ml or 10.4 mg/kg. Because the drug is concentration dependent, the dose is administered once daily. For amikacin, the targeted PDC would be 8 µg/ml X 10 = 80 µg/ml. The dose would be 80 µg/ml X 0.23 L/kg = 18 mg/kg. Interval. Intervals are not as likely to be as flexible as doses. For time dependent drugs, the recommended interval should not be prolonged; shorter intervals might be considered in problematic infections. For concentration dependent drugs, once daily dosing is likely to be effective. Currently, there is no data to support a shorter dosing interval to be problematic; however, consideration should be given to maintaining a proper inhibitory quotient for each dose. The route of administration becomes particularly important for life threatening infections, or infections in tough to penetrate tissues. For such infections, an IV route is indicated at least initially to enhance drug concentrations at the site. COMBINATION ANTIMICROBIAL THERAPY Indications and Contraindications Combination antimicrobial therapy can be a powerful tool in the treatment of bacterial infections. For example, rational combination antimicrobial therapy may be the single most effective action taken to enhance antimicrobial efficacy in the chronic or serious infection. Indications for combination therapy include: achieving a broad antimicrobial spectrum for empirical therapy; treatment of a polymicrobial infection involving organisms not susceptible to the same drugs; reducing the risk or advent of antimicrobial resistance; and reducing the risk of adverse drug reactions by minimizing exposure to potentially toxic antimicrobials (Schimpff, 1989; Kapusnik, 1989; Neu, 1994). Combination therapy often is used when dealing with infections caused by both an aerobic gram positive and aerobic gram negative organism, or an aerobic and anaerobic infection (Kapusnik, 1989; Neu, 1984; Boothe, 1990; Nostrandt, 1990). Polymicrobial therapy may be best exemplified by the treatment of peritonitis. Drugs effective against organisms causing infection early in the pathyophysiology, Enterobacteriaceae (eg, aminoglycosides, fluorinated quinolones) should be combined with drugs effective against anaerobes (eg, penicillins, metronidazole or clindamycin) which complicate infection by causing abscessation in later stages. Combination of metronidazole with either gentamicin or ciprofloxacin appears to be equally effective in preventing infection following abdominal trauma (Tyburski 1998); metronidazole when combined with ciprofloxacin is effective as a perioperatove antibiotic in colorectal surgery (Rohwedder 1993) and appears equal in efficacy to impipenem/cilastin for the treatment of complicated intraabdominal infections (Solomkin 1996). The combination of ciprofloxacin and metronidazole has proven effective in the treatment of Crohn’s disease (Prantera 1996, Greenbloom 1998). Triple antibiotic therapy has been used to sterilize in humans to sterilize infected root canals (Sato 1996). Combination therapy is not always indicated for the treatment of polymicrobial infections. For example, eradication of one organism may allow host mechanisms to control or eradicate other organisms present in infection. Additionally, the advent of newer antibiotics whose spectrum includes multiple classes of microorganisms (eg, imipenem) may preclude combination therapy. The ongoing concern regarding global resistance of microbes to antimicrobial drugs may justify this indication in order to minimize antimicrobial resistance in the individual patient (Jones 1997). Individual drugs in the combination should be selected in part based on different mechanisms by which microbes develop resistance to the drugs. Avoidance of toxicity by combining antimicrobials was first exemplified by triple sulfonamide drugs: the use of three drugs prevented any one drug reaching concentrations sufficient to cause nephrotoxicity. The relative safety of antimicrobials reduces the need for combination therapy in order to avoid toxicity; a notable exception occurs for aminoglycoside-induced nephrotoxicity. Combination therapy should not be used with the intent of using a lower dose of the toxic drug. Such an approach increases the risk of nontherapeutic concentraitons being achieved for the potentially-toxic drug, thus increasing the risk of resistance. Although minimum inhibitory concentrations may be lower for combined drugs compared to individual drugs (Oliver 1999), the advantage of combination may be lost if inappropriately decreased doses are used. Combination therapy should be undertaken with the intent of decreasing the duration of antimicrobial exposure by enhancing efficacy and minimizing resistance; full doses of both antibiotics should be used. The primary reasons to avoid combination therapy include increases in the risk of supra-infection, risk of toxicity (if both drugs are potentially toxic), cost, and inconvenience to the patient (Kapusnik, 1989, Maertens 1998). Sequelae of Combination Antimicrobial Therapy Antimicrobials to be used in combination therapy should be selected rationally and should be based on target organisms as well as mechanisms of action that complement one another (Figure 1). Combinations might result in a number of sequelae which are defined by in vitro methods. The most common method used to evaluate the effects of combined drugs is the checkerboard method. This method generally tests concentration of each antimicrobial that range from four to five dilutions below to two dilutions above the minimum inhibitor concentration (MIC) of each antimicrobial. (Eliopoulos, 1996). Drugs are usually added to Mueller-Hinton broth. Efficacy is evaluated by comparing the fractional inhibitory concentration (FIC; the MIC of the drug when combined divided by the MIC of the drug when used alone) of the drugs (den Hollander 1998). Care must be taken in predicting or overinterpreting the sequelae of antimicrobial combinations based on this method applicability to clinical situations may be difficult to establish. In addition, the combined effects of two or more antimicrobials is likely to differ with the organism targeted by the combination. The system more effectively detects positive interactions between two antimicrobials; the following interactions can be defined by this system. Autonomy or indifference occurs when results of the two drugs combined do not differ significantly from the results obtained with the most effective drug studied alone (FICA + FICB = FICA or FICB). Additivity occurs when the results of the two drugs equals the sum of the results for each of the drugs used separately (FICA + FICB = 1.0) where as antagonism occurs when the result of the two drugs combined is significantly less than the additive response (FICA + FICB = > 1.0). Synergism occurs when the results of the two drugs combined is significantly greater than the additive response ((FICA + FICB = < 0.5). Other methods have been developed to evaluate combination antimicrobial efficacy in order to overcome the disadvantages of the checkerboard method. Examples include agar diffusion methods, killing curves, which measure bactericidal activity (and therefore may be more clinically relevant than the checkerboard method), serum bactericidal testing (to assure that antimicrobial concentrations tested in the former methods are clinically relevant) and kinetic spectrophotometric methods which offer a better evaluation of the dose-response relationship. Each method is associated with advantages and disadvantages; all are qualitative and none have been validated in clinical patients. For "bactericidal" drugs (drugs for which the MIC is close to the minimum bactericidal concentration [MBC], the checkerboard method is probably most appropriate; for drugs whose MBC and MIC are not close ("bacteriostatic " drugs), the killing curve methodology is probably most clinically applicable (Eliopoulus 1996). Antagonistic Combinations Avoidance of antagonism is particularly important to patients with inadequate host defenses (Schimpff, 1989; Kapusnik, 1989; Neu, 1994; Eliopoulos, 1991). In general, "bacteriostatic" drugs which inhibit ribosomes and thus microbial growth (eg, chloramphenicol, tetracyclines, and erythromycin) should not be combined with drugs whose mechanism of action is dependent upon protein synthesis, eg, growth of the organism (eg, beta-lactams) or formation of a target protein. An antagonistic combination is best exemplified by the combination of a ?-lactam antimicrobials (eg, penicillin) and inhibitors of ribosomal activity (eg, dtetracyclines, chloramphenicol); antagonistic effects have been well documented for several combinations. The bactericidal activity of and continued degradation or destruction of the microbial target of ?-lactams is dependent on continued and active cell growth. As the organism grows, its cell wall is destroyed and replaced with new, albeit compromised, cell wall. If the organism is not rapidly growing, cell wall degradation and synthesis will not continue, and the beta-lactam will not be incorporated into the cell wall. Inhibitors of ribosomal activity preclude beta-lactam bactericidal activity by preventing rapid cell growth. (Jawetz 1951a). Although this effect has been documented for some organisms, antagonism can not always be documented with these drug combinations. For example, chloramphenicol does not appear to antagonize the bactericidal effects of penicillin against Streptotoccus pneumoniae (Bodine 1977). Ribosomal inhibitors also may inhibit autolysins, enzymes responsible for microbial cell wall degradation (Tomasz 1975). The degree of antagonism between fluorinated quinolones and growth inhibitors is controversial; antagonism has been reported with use of ciprofloxacin and chloramphenicol (Eliopolous, 1991), but impaired efficacy was not detected in other studies (Zeiler, 1987). Ribosomal inhibitors also antagonize the effects of aminoglycosides, which also are ribosomal inhibitors (Jawetz 1951b). The mechanism of this antagonism is not clear, but may reflect inhibition of active transport mechanism upon which aminoglycoside activity is dependent (Eliopoulus 1996). Antagonism between chloramphenicol and gentamicin in particular has also been documented (Eliopolous 1991). Interestingly, inhibition of protein synthesis by a ribosomal inhibitor may also facilitate the actions of another antimicrobial (see synergism). Antagonism also may occur if two drugs that act at the same site are combined with one another, resulting in competition for the same binding site. Antagonism by this mechanism is exemplified by the combination of two or more drugs which target the 50s subunit of ribosomes (macrolides, lincosamides and chloramphenicol) (Weisblum 1967). Combinations of drugs active at the same ribosomal site can result in simultaneous resistance to both drugs (Barber 1964). Clinically, evidence of antagonism among 50s ribosomal subunit inhibitors has not been documented. The combination of two or more beta-lactams can similarly result in antagonistic effects. Combinations of beta lactams containing one agent that induce beta-lactamase production (eg, selective cephalosporins, particularly cefoxitin) with another beta-lactam can result in antagonistic actions (Gootz 1982). Beta-lactams which target the same penicillin-binding proteins also can result in antagonistic actions. Chemical antagonism is also possible between two or more antimicrobials (Zeiler 1987; King, 1989). Aminoglycosides or fluorinated quinolones (weakly basic drugs) are chemically inactivated by penicillins (weakly acidic drugs). Inactivation prevents the combination of these drugs in intravenous fluids. However, chemical inactivation is unlikely at concentraions generally achieved with therapeutic use. Exceptions may occur as once-daily therapy at high doses becomes the standard regimen for the aminoglycosides and fluorinated quinolones. Plasma concentrations also may become sufficiently high for chemical inactivation to occur in patients with impaired renal function sufficient to compromise renal excretion of the drugs. In human medicine, measurement of aminoglycoside concentrations is recommended to avoid inactivation (Davies 1975) at sufficient concentrations. Ticarcillin has been used therapeutically to reduce the risk of toxicity in a patient overdosed with an aminoglycoside (Olin, 1992). Chemical antagonism is unlikely in most clinical uses of these drugs. However, the risk of antagonism is increased with simultaneous IV use of high doses of both drugs such as might occur if aminoglycosides are administered once daily. Potential chemical interactions between other antimicrobials should be identified prior to combination therapy. Certainly, antimicrobials should not be mixed in the same syringe or IV line unless the lack of antagonism has been confirmed (King, 1989) unless studies support otherwise. An example of an apparently acceptable combination is ciprofloxacin with metronidazole for combined efficacy in treatment of serious abdominal infections (Messerschmidt 1988). Additive Combinations Generally, drugs that have the same mechanism of action probably act in an additive fashion. For example, chloramphenicol and clindamycin bind the same 50s ribosomal subunit and will antagonize one another, but since tetracyclines bind to the 30s subunit, combination of either drug with tetracyclines conceivably might be additive (although this has not been studied). Additive effects probably occur when active metabolites are produced from an active parent compound, such as metabolism of enrofloxacin to ciprofloxacin (Kung, 1993). However, antagonistic effects might occur if the drugs compete for a limited number of target sites (ie, chloramphenicol and erythromycin). In contrast, synergistic actions might occur if the antimicrobial targets are subtly different. For example, combination of different ?-lactams generally results in additive antimicrobial activity. However, if the two antimicrobials target different penicillin-binding-proteins, their combined effect may actually be synergistic ("double ?-lactam therapy") (Hopefl, 1991; Pedler, 1984). Double –beta lactam therapy may also be of benefit if one drug protects the other against destruction by beta-lactamases. Such is the rationale for combining clavulanic acid or sulbactam with some beta-lactams; however other beta-lactam combination have are probably effective (eg, aztreonam protection of cefepime) (Lister 1998) In contrast, combinations of other ?-lactam antibiotics (including combined cephalosporins) are antagonistic (Pedler, 1984). The different sequela of combined ?-lactam therapy might be in the penicillin binding proteins targeted by each drug. Synergistic Combinations Synergism between antimicrobials is most likely to occur if the two antimicrobials kill bacteria through independent mechanisms (eg, an aminoglycoside coupled with a beta-lactam) or through sequential pathways towards the same target (Richards, 1993; Eliopoulos, 1991) (Figure 2). Four well established mechanisms of synergism exist. 1. Sequential inhibition of a common pathway is best exemplified by the combination of two bacteriostatic drugs: a sulfonamide coupled with either trimethoprim or ormetoprim (Bushby 1973). Both drugs target bacterial synthesis of folic acid, but each through a different enzyme. Their combined effect results in bactericidal activity. inhibition of an enzyme of destruction (eg, beta-lactamase), 2. Drugs which sequentially inhibit cell wall synthesis are best exemplified by a penicillin combined with vancomycin (Elipolous 1996). 3. Inhibition of a destructive enzyme, best exemplified by the combination of clavulanic acid with amoxicillin (Reading 1977). The clavulanic acid "draws the attention" of Figure 2. Mechanisms of antimicrobial action should be the basis for initial selection of combination therapy. Drugs which inhibit ribsomal inhibitors should be used cautiously with other antimicrobials. the beta-lactamase, allowing the other beta-lactam antibiotic (amoxicillin) to bind to transpeptidase enzymes necessary for successful cell wall synthesis. Interestingly, drugs which inhibit ribosomal activity and subsequently protein synthesis can facilitate the actions of another antibiotics. For example, chloramphenicol can synergize the activity of some penicillins against organisms resistant to beta-lactams due to beta-lactamase production apparently by inhibiting beta-lactamase activity (Luboshitzky 1973). This apparently contradictory effect (see antagonism) occurs at low concentrations of chloramphenical; higher concentrations result in the traditionally recognized antagonism. Clindamycin may cause a similar synergistic effect in resistant beta-lactamase organisms (Sanders 1973). 4. Facilitation of entry of an antimicrobial through the cell wall. Synergism between ?-lactams and aminoglycosides exemplifies synergism as a result of enhanced penetration of an antimicrobial (Moellering 1971) (Figure 2). The aminoglycoside is not able to penetrate the organism (and thus reach its ribosomal targets) until the beta-lactam results in a cell wall that is permeable to the former drug. Although synergism is expected because the mechanisms of action of the aminoglycoside and beta-lactam compliment one another, efficacy is enhanced further because aminoglycoside movement into the bacteria is enhanced by increased cell wall permeability induced by the ?-lactam (Eliopoulos, 1991). This synergistic effect has been documented in a variety of organisms (gram negative and gram positive). Enhanced movement of extracellular drug into the microbe and may explain how synergism between an aminoglycoside and a beta-lactam occurs in Bacteroides melaninogenicus (anaerobic) organisms (Eliopoulus 1996). Aminoglycosides, dependent on oxygen for active transport into the anaerobic microbe, generally are not effective against obligate anerobes. Enhanced movement in a bacteria may occur for other drugs (eg, potentiated sulfonamides) when combined with ?-lactams. Other synergistic antimicrobial combinations have been described, although the mechanism is not known. For example, clindamycin, a drug whose spectrum does not include Pseudomonas aeruginosa acts in a synergistic fashion with aminoglycosides against the organism, although only at concentrations of clindamycin much higher than those generally achieved at recommended doses (Leng 1975). Synergism between these two drugs has been documented in rat models of peritonitis at clinically relevant doses (Louie 1977). Enrofloxacin, a drug whose anaerobic spectrum is weak acts synergistically with metronidazole against a number of anaerobic spectrum (unpublished data, Bayer Animal Health, 1999). Synergism can result in efficacy against an organism resistant to both drugs used individually. For example, polymixin B when combined with a sulfondamide - trimethoprim combination act synergistically against Proteus and Serratia sp. (Greenfield 1970). Ciprofloxacin, a drug characterized by a poor spectrum of anaerobic activity, enhances the efficacy of metronidazole against selected anaerobic organisms (Werk 1988). Synergism has been documented with fluorinated quinolones combined with aminoglycosides against Pseudomonas aeruginosa, but in less than 1/3 of the organisms tested. The combination of a beta lactams with aminoglycosides more consistently induces synergistic effects, although the range of 10-70% suggests caution when assuming synergism (Elipolous 1996). A synergistic effect has been documented in the treatment of gingivitis and periodontal disease in both dogs and cats when metronidazole is combined spiramycin (Weideman 1992, Hennet 1991, Feik 1991). ANTIMICROBIAL PROPHYLAXIS The prophylactic use of antibiotics must be distinguished from treatment. The presence of infection, or anticipated infection following bacterial contamination (ie, a compound fracture; contamination of abdominal contents with intestinal fluid) indicates the need for treatment rather than prophylaxis. If antimicrobial prophylaxis is to be implemented in anticipation of an invasive procedure (ie, surgery), the following should serve as a basis for selection: the antimicrobial should target the most likely pathogenic organism; adequate concentrations of drug should be at the site of invasion prior to potential contamination; the antimicrobial should either have a long elimination half-life, or be redosed during lengthy procedures; the least toxic drug should be selected; and the duration of therapy should be as short as possible (Schimpff, 1989; Neu, 1994). Prophylactic therapy should not be used indiscriminately in the immunocompromised animal. The granulocytopenic patient is particularly predisposed to the development of suprainfection. Suprainfection occurs in 10 to 20% of human granulocytopenic patients receiving empirical broad spectrum antimicrobials. Prolonging therapy increases the chance that suprainfection will occur (Schimpff, 1989). Prophylactic suppression of gastrointestinal flora is recommended in human patients that are profoundly granulocytopenic for more than two weeks. Traditional use of non-absorbable antimicrobials effective against aerobic gram-negative organisms (eg, neomycin), and drugs which target anaerobic organisms (eg, metronidazole), are being replaced by use of trimethoprim-sulfonamide combinations or fluorinated quinolones (Schimpff, 1989). Trimethoprim/sulfonamide combinations are more palatable and cheaper, yet they are equally effective in preventing infections when compared to more expensive drugs in human critically ill patients. Fluorinated quinolones allow persistence of anaerobic organisms in the gastrointestinal tract, thus reducing overgrowth of resistant gram-negative organisms and preventing rapid repopulation and overgrowth of aerobic gram-negative organisms as the antimicrobial is discontinued. Drug ENR DIF ORB MAR Dose (mg/kg) 2.5 10 2.5 2.5 Serum 1.2 (2.4)* 1.2 2.3 2.0 (at 1.5 hrs) CSF/Brain 0.25 (0.5)* NR NR NR Prostate 1.4 (2.8)*   NR 5.6 Skin 0.7 (1.4)* NR NR 1.9 Bone 2.1 (4.2)* (marrow) 7.2 NR 3.1 (marrow) Table 1. Comparative data provided on PFL for the flourinated quinolones (FQ). MIC90 data from enrofloxacin (ENR) represents the most recent NCCLS report (to be on an updated PFL). The time of collection of tissue data (1 hour for enrofloxacin, 2 hours for others) is similar among drugs, although time to peak concentrations are likely to very among drugs and this data does not necessarily represent peak concentrations for any drug. The dose differs among the drugs: difloxacin (DIF) data represents the higher end of its recommended dosing regimen compared to the lower end of the dosing regimen for orbifloxacin (ORB) and marbofloxacin (MAR). The data for ENR was collected at half the current recommended lower end of the dosing regimen; and extrapolated number in parenthesis represents approximate concentrations at the lower end of the dosing range. Not all data has been made available for all tissues and each drug. This data is probably most relevant for treatment of problematic infections (see text). To treat a prostatic infectioin caused by E.coli, using the MIC90 data and tissue concentrations, a dose of 5 mg/kg of ENR will generate (in normal tissue) a ratio of tissue drug concentration to MIC90 of approximately 22 (2.8 / .125); the same ratio for DIF is 12 (at the high end of the dosing range) and for MAR, 60. Tissue data is not available for ORB, but its MIC90 is at least 3 times that of the other drugs, suggesting a ratio that is lower than that of the other drugs. The inhibitory quotient indicates the lower dose may be appropriate for ENR and MAR, but the higher dose is necessary for DIF. The proper dose can not be assessed for ORB. Since the FQ are concentration dependent (at least for gram negative organisms), a 24 hour dosing interval is appropriate for each. ORG ENR DIF ORB MAR Staph 0.25 0.46 0.39 0.25 E coli 0.06 0.11 0.39 0.06 Klebsiell a 0.12 0.11 NR 0.06 Proteus sp 0.25 1.83 NR 0.125 Pseudom on 1 0.92 (n=5) 12.5 NR Table 2. Selected examples of combinations proven to be additive or synergistic based on in vitro data Organism Drug 1 Drug 2 Nocardia TMP Sulfonamide   Clavamox Cefotaxime   TMPS/sulfonamide Imipenem   TMP/sulfonamide Amikacin Bacteroides Metronidazole Clindamycin   Aminoglycoside Beta-lactams Myuobacterium Ethambutol Rifampin, aminoglycosides, clarithromycin, or fluorinated quinolone Staphyloccocus aureus Clavulanic acid Many penicillins   Penicillins Aminoglycosides   Penicillin G Chloramphenicol (bacteriostatic)   Imipenem, aminoglycosides Vancomycin Eschericia coli, Klebsiella Ampicillin, penicillin or cephalothin Dicloxacillin Pseudomonas aeruginosa Methicillin Ampicillin   Clafulanic acid Nafcillin or ampicillin   Vancomycin Cephalothin   Beta lactam Aminoglycoside   Clindamycin Aminoglycoside [back to top]